Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The enzymes of the cytochrome P-450 (P-450)³ superfamily are involved in oxidative metabolism of xenobiotics in several animal species. In a recent review, identification of 481 genes and 22 pseudogenes of the P-450 superfamily were cited in 85 eukaryotic and 20 prokaryotic species (Nelson et al., 1996). These enzymes of the P-450 family are responsible for the metabolism of a wide variety of xenobiotic and endogenous compounds (Rendic and DiCarlo, 1997). The CYP3A isozymes of the P-450 superfamily have been detected in several species including humans, rats, rabbits, guinea-pigs, sheep, primates, dogs, goats, and miniature pigs (Nelson et al., 1996).

Porcine hepatocytes continue to find increasing application in bioartificial liver devices (Gerlach, 1996; Joly et al., 1997; Watanabe et al., 1997). Our group has developed a collagen gel-entrapment bioartificial liver containing porcine hepatocytes (Hu et al., 1997). The device was shown to be effective in a D-galactosamine liver injury model in canines (Sielaff et al., 1995b). One potential method to improve the function of the bioartificial liver is to increase the biotransformation capacity of the cells via induction. Very little is known about P-450-mediated metabolism in pigs, or in cultured porcine hepatocytes, although there have been some reports suggesting the presence of CYP3A isoforms in cultured porcine hepatocytes (Witkamp et al., 1995) and in pig livers (van den Broek et al., 1984).

In humans and in rats, the CYP3A isozymes constitute about 25 to 30% of all the P-450 forms expressed in the liver (Imaoka et al., 1988; Shimada et al., 1994). In humans, the CYP3A family of enzymes are responsible for the metabolism of many drugs and putative carcinogens (Li et al., 1995a). The induction and expression of these enzymes is therefore of great interest to investigators studying xenobiotic metabolism in different species.

CYP3A isoforms respond to induction in a species-dependent manner. Kocarek et al. (1995) studied the induction of CYP3A isoforms in hepatocytes isolated from rat, rabbit, and human sources. The induction of CYP3A isoforms in cultured hepatocytes was quantitated based on levels of CYP3A mRNA and immunoreactive protein. Rifampin (10 µM) was a potent inducer of CYP3A isoforms in rabbits and humans, and a weak inducer in rats. Dexamethasone (10 µM) was a potent inducer in rats and rabbits, and a moderate inducer in humans. Pregnenolone 16-carbonitrile (10 µM) was a potent inducer of CYP3A in rat hepatocytes, and a moderate inducer in human hepatocytes, but failed to induce CYP3A in rabbit hepatocytes (Kocarek et al., 1995). More recently, Olsen et al. (1997) reported the induction of testosterone-6-hydroxylation (CYP3A enzyme activity) in cultured pig hepatocytes upon treatment with rifampin, carbamazepine, and phenobarbital but not dexamethasone. The above studies indicate a pronounced species dependence in the response of CYP3A isozymes to induction.

It has been shown that midazolam (MDZ) is a selective substrate of CYP3A, and is metabolized to 1'-hydroxymidazolam (1'-OH MDZ), 4-hydroxymidazolam (4-OH MDZ), and 1',4-dihydroxymidazolam (1',4-diOH MDZ), in human and rabbit liver microsomes (Fig. 1) (Fabre et al., 1988a; Gorski et al., 1994; Ghosal et al., 1996). Hence MDZ could be used as a substrate probe to characterize CYP3A induction in porcine hepatocytes, without the interference of alternate metabolic pathways. The present investigation was carried out to characterize metabolism of MDZ in porcine hepatocytes, and to determine the inducibility of CYP3A isoforms in porcine hepatocyte cultures. The inducibility of CYP3A isoforms was evaluated by measuring the metabolism of MDZ, a known substrate for CYP3A4 and CYP3A5 in humans. In addition, the induction was also evaluated at the molecular level to ascertain the mechanism of induction.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Metabolism of MDZ by isozymes of the CYP3A subfamily.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials. Male Dorac pigs (8-12 kg) were obtained from Midwest Swine Research (Gibbon, MN). Collagenase D was obtained from Boehringer Mannheim (Indianapolis, IN). Primaria and other cultureware were obtained from Becton Dickinson Laboratories (Franklin Lakes, NJ). Williams E media, L-glutamine, and HEPES were purchased from Life Technologies (Grand Island, NY). Penicillin and streptomycin were obtained from Celox (Hopkins, MN). Bovine-porcine pancreas insulin was purchased from Eli-Lilly (Indianapolis, IN). All other media components were purchased from Sigma Chemical Co. (St. Louis, MO). Type I collagen (Vitrogen) was obtained from Collagen Corp. (Palo Alto, CA). MDZ, 1'-OH MDZ, 4-OH MDZ, 1',4-diOH MDZ, and Ro 22-0991/000 (internal standard) were a gift from Dr. Ernest Bush at Hoffman-LaRoche, Inc. (Nutley, NJ). The anti-CYP3A antibody was a generous gift from Kenneth E. Thummel, College of Pharmacy, University of Washington (Seattle, WA). The goat anti-rabbit IgG linked to alkaline phosphatase was obtained from Sigma Chemical Co. (St. Louis, MO). The 46-mer cDNA oligonucleotide probe was synthesized by Oligos Etc., Inc. (Wilsonville, OR). The substrate system for the alkaline phosphatase was obtained from Bio-Rad (Hercules, CA). The Micro-FastTrack kit for RNA isolation was obtained from Invitrogen Corp. (San Diego, CA). High purity grade tetrahydrofuran and t-butyl methyl ether were obtained from Burdick and Jackson (Muskegon, MI). All other chemicals and solvents were of analytical grade. Membranes for Western and Northern blotting were obtained from Schleicher & Schuell (Keene, NH).

Preparation of Collagen Gels. Collagen gels were prepared in six-well Primaria culture plates and 60-mm plastic dishes. A solution of collagen was prepared by mixing three volumes of type I collagen and one volume of 4-fold concentrated Williams E media supplemented with 0.2 U/liter bovine-porcine insulin, 100 U/liter penicillin, 100 mg/ml streptomycin, 2 mM L-glutamine, and 15 mM HEPES at 4°C. The pH of the solution was adjusted to 7.4 by addition of a sterile solution of 0.1 N NaOH. The collagen solution (0.5 ml) was added to each well of the six-well plates, and 1 ml was added to the culture dishes to obtain a final protein concentration of 112.5 µg/cm².

Hepatocyte Culture and Induction. Porcine hepatocytes were isolated from high-health-status male Dorac pigs (10 kg) by the method reported by Sielaff et al. (1995a). Routine hepatocyte viability at the time of isolation was >92% as determined by trypan blue dye exclusion. The hepatocytes were plated in six-well collagen coated plates at a cell density of 0.5 × 10⁶ cells/well, and in the 60-mm culture dishes at a density of 3 × 10⁶ cells/dish. The isolated hepatocytes were cultured in hormonally defined culture media composed of Williams E media supplemented with 100 U/ml penicillin, 100 µg/ml streptomycin, 0.2 U/ml insulin, 1 nmol/ml dexamethasone, 4 ng/ml glucagon, 25 ng/ml epidermal growth factor, 50 ng/ml linoleic acid, 20 ng/ml Gly-His-Lys (liver cell growth factor), 6.25 µg/ml transferrin, 500 µg/ml albumin, 0.1 µM CuSO₄ 5H₂O, 3 nM H₂SeO₃, 50 pM ZnSO₄ 7H₂O, and 15 mM HEPES at pH 7.4.

Use of MDZ as a Substrate Probe. MDZ was used as a selective substrate probe to measure the activity of the CYP3A isozyme from the P-450 superfamily of enzymes. After induction of the porcine hepatocytes for 72 h, the culture media was replaced with fresh media containing MDZ (50 µM), and either rifampin (50 µM), dexamethasone (50 µM), or DMSO (0.1% v/v). Media samples were collected at 0, 5, 15, 30, 45, 60, and 120 min after the addition of MDZ. The samples were frozen at 80°C until they were analyzed by HPLC.

Extraction and HPLC Analysis of MDZ and its Metabolites from Culture Media. The analysis of MDZ and its metabolites involved liquid-liquid extraction, followed by quantitative HPLC analysis. Media (0.5 ml) was mixed with 25 µl of an internal standard solution (RO 22-0991, 500 µg/ml) and 4 ml of t-butyl methyl ether in a screw-capped tube and mixed for 5 min on a reciprocating platform shaker. The tubes were then centrifuged for 5 min at 2500 rpm in a Baxter Multifuge centrifuge (Baxter, McGaw Park, IL). The aqueous layer was frozen by placing the tubes in a dry ice- acetone bath and the organic layer was transferred to 13 × 100 culture tubes. The organic layer was evaporated under N₂ at 50°C, and reconstituted in 200 µl of mobile phase. Standard samples were prepared in a similar fashion from media containing known concentrations of MDZ and its metabolites. The standard solutions of MDZ and metabolites were prepared in methanol, and 25 µl of these solutions were added to 0.5 ml of media. Standard curve concentrations ranged from 100 to 0.25 µM for MDZ, from 50 to 0.125 µM for 1'-OH MDZ and 4-OH MDZ, and from 12.5 to 0.03125 µM for 1',4-diOH MDZ.

Preparation of Microsomes from Hepatocytes in Culture. Microsomes were prepared from porcine hepatocytes in culture by a modification of the method reported by Schuetz et al. (1988). Hepatocytes cultured under control culture conditions and under rifampin-induced culture conditions (culture conditions described earlier) were used in the preparation of microsomes. Briefly, the hepatocytes (control and induced cultures) from five plates were washed with ice-cold PBS three times, removed from the plates by gentle scraping, and homogenized in 0.01 M phosphate buffer (pH 7.4) containing 1.15% KCl. The cells were homogenized by three strokes of a Potter-Elvehjem homogenizer. After homogenization, the lysates were centrifuged at 9000g for 20 min at 4°C. The supernatant was collected and spun at 100,000g for 60 min at 4°C. The supernatant was removed and the pellet was resuspended in 1 ml of 20% glycerol/80% 0.1 M phosphate buffer (pH 7.4) containing 1.0 mM EDTA for storage at 80°C. Microsomes were also prepared from pig livers by a similar method with an additional pyrophosphate buffer washing step (Remmel and Sinz, 1991). For the in vivo induction, a 10-kg Dorac pig was given an oral suspension of rifampin compounded in Ora-Sweet and Ora-Plus (Paddock Labs, Minneapolis, MN) at dose of 20 mg/kg/day for 4 days. Microsomes from the rifampin-treated pig was prepared by a method identical with that from untreated pigs.

Western Blot Analysis. Microsomal proteins (40 µg) from rifampin-treated and control porcine hepatocyte cultures and from untreated and rifampin-treated pigs were resolved on 9% SDS-polyacrylamide gels (15 × 15 cm/1 mm width) with a stacking gel composed of 3% acrylamide. Samples were electrophoretically transferred to nitrocellulose membranes on a Bio-Rad semidry transfer apparatus (Bio-Rad, Hercules, CA) according to instructions provided by the manufacturer of the apparatus. The blots were developed with a primary antibody raised against human CYP3A4 followed by a goat anti-rabbit IgG linked to alkaline phosphatase. The individual blots were visualized with a color-developing system (BioRad substrate system). Microsomal proteins from untreated pig livers and rifampin-treated pig livers were used as controls.

Isolation of mRNA and Northern Blotting. Total RNA from the hepatocytes was isolated from five culture dishes by the method of Chomczynski and Sacchi (1987). Messenger RNA was isolated from the total RNA with a Micro-FastTrack kit obtained from Invitrogen (San Diego, CA). The isolated mRNA samples were resolved on denaturing 1.5% agarose gels containing 2 M formaldehyde. Uniform loading of mRNA samples was monitored by ethidium bromide staining of the agarose gels and visualization of the RNA bands before transfer of RNA to the membrane. The resolved RNA samples were then transferred by capillary action onto a nylon membrane (Schleicher & Schuell, Keene, NH). The membrane was incubated overnight at 52°C in a prehybridization solution composed of 5× saturated sodium citrate (SSC; 1× SSC: 0.15 M NaCl, 0.015 M Na Citrate), 2.5× Denhardt's reagent (1× Denhardt's Reagent: 0.02% Ficoll, 0.02% polyvinyl pyrrolidone, 0.02% BSA), 15 mM sodium phosphate buffer (pH 7.2), 1 µg/ml salmon sperm DNA, and 0.1% SDS. The membrane was then hybridized overnight at 52°C with a radiolabeled cDNA oligonucleotide probe (described later). The hybridization solution contained 50% formamide, 5× SSC, 2× Denhardt's reagent, 15 mM sodium phosphate (pH 7.2), 5 µg/ml salmon sperm DNA, 0.1% SDS, and 10% dextran sulfate. After hybridization, the membranes were washed two times in 6× SSC, 0.1% SDS for 15 min at 45°C, and two times in 2× SSC, 0.1% SDS for 10 min at room temperature. The final wash was in 2× SSC, 0.1% SDS for 15 min at 45°C. The membranes were then exposed to X-ray film for 3 days at 80°C.

Oligonucleotide Probe for Northern Blot. The cDNA oligonucleotide probe (46-mer) was designed based on a 17-amino-acid sequence around the conserved Cys residue of CYP3A1 in the heme binding region (Strömstedt et al., 1991). This sequence has less than 50% homology to other rat P-450s and is highly conserved in the sequences of CYP3A mRNA in several species. The oligonucleotide probe was radiolabeled at the 5'-end to >10⁸ cpm/ng with a nick-translation kit (BRL, Gaithersberg, MD). The sequence of the probe was:

Quantitation of RNA and Proteins on Blots. The autoradiographs (from the Northern blots) and the membranes (from the Western blots) were scanned on a GS-700 imaging densitometer (Bio-Rad, Hercules, CA) and quantitated with Molecular Analyst software Version 2.1 (Bio-Rad). All levels were corrected for background signal during quantitation.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of Rifampin and Dexamethasone on MDZ Metabolism in Cultured Porcine Hepatocytes. Treatment of cultured porcine hepatocytes with rifampin led to a significant increase in the metabolism of MDZ when compared with control cultures (Table 1). No difference was observed in the formation rates of the metabolites of MDZ between the dexamethasone-induced cultures and the control cultures. All cultures produced 1'-OH MDZ as the major metabolite. In addition, 4-OH MDZ was also formed in the induced cultures. Table 1 shows the rates of formation of the different metabolites of MDZ under induced and control conditions. Rifampin treatment resulted in a significant 6-fold induction in the rate of formation of 1'-OH MDZ (p < .01) and a 9-fold induction in the formation of 4-OH MDZ (p < .01). The metabolic profiles for 1'-OH MDZ formation in dexamethasone- and rifampin-induced cultures are shown in Fig. 2.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Profiles for 1'-OH MDZ formation in rifampin (A) and dexamethasone (B) induction experiments. (), induced cultures; (), control cultures. The values are means from experiments performed on hepatocytes from three different pigs, and the error bars are S.D.

Effect of Rifampin on CYP3A mRNA Synthesis. Northern blot analysis revealed the presence of one band corresponding to ~2.4 kB (Fig. 3). This band was seen in the lanes containing mRNA isolated from control cultures, and from rifampin-induced cultures. The intensity of the band in the lane with the mRNA from rifampin-induced cultures was roughly 10- to 12-fold greater than the intensity of the band in the lane for mRNA from control cultures, as determined by transmittance densitometry. The corrected values for the band intensities for the bands from the rifampin-treated cultures and control cultures were 5.189 and 0.383 arbitrary units/mm², respectively.

View larger version (60K): [in this window] [in a new window]   Fig. 3.   Northern blot from rifampin-treated and control hepatocyte cultures. The approximate size of the band was 2.4 kB.

Effect of Rifampin on CYP3A Protein Expression in Porcine Hepatocytes. Western blot analysis with an anti-CYP3A antibody indicated the presence of four bands of immunoreactive proteins in pig liver microsomes (Fig. 4). The CYP3A bands were seen in the molecular mass range from 51 to 55 kDa, as determined from the molecular weight markers. Only two bands (Bands 2 and 3, 52 and 51.5 kDa, respectively) were detected in all lanes. In microsomes prepared from pigs pretreated with rifampin, these two bands were significantly increased compared with microsomes from untreated pigs. These two bands were also detected in microsomes prepared from either control or induced porcine hepatocytes. The intensities of the bands were determined by reflectance densitometry, and the values obtained from the densitometry experiments are shown in Table 2. There was a greater than 10-fold increase in the levels of a CYP3A immunoreactive protein (Band 2) and a 6-fold increase in the levels of another immunoreactive protein (Band 3) in microsomes from rifampin-induced hepatocytes compared with uninduced controls.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Western blot on microsomal proteins isolated from porcine hepatocyte cultures. Lanes A and B were loaded with 20 and 40 µg microsomal protein from control pig livers. Lanes C and D were loaded with 10 and 20 µg of liver microsomal protein from rifampin-treated pigs. Lanes E and F were loaded with 20 and 10 µg of microsomal proteins from control cultures and rifampin-induced cultures, respectively. Lane G was loaded with 0.2 pmol of purified human CYP3A4 protein (positive control). The bands range in molecular mass from 55 to 50 kDa. No other immunoreactive proteins were seen on this blot. (Markers not shown).

Concentration Dependence of the 4-OH/1'-OH MDZ Ratio. In all culture conditions tested, we observed that the major metabolite was the 1'-OH MDZ. Ochs et al. (1987) reported that the only circulating metabolite observed in the plasma of 17- to 37-kg domestic pigs was 4-OH MDZ after i.v. or oral dosing. At the high concentration (50 µM) of MDZ, 1'-hydroxylation predominated over 4-hydroxylation in hepatocyte incubations. Because the report of Ochs et al. (1987) appeared to contrast with our hepatocyte results, a small pilot experiment was conducted with microsomes prepared from a rifampin-treated pig and an untreated 10-kg Dorac pig. Comparison of peak area ratios by HPLC revealed that the ratio of 4-OH/1'-OH MDZ increased from 0.99 ± 0.04 at 50 µM MDZ to 1.53 ± 0.34 at 0.5 µM in untreated liver microsomes. In microsomes from a rifampin-treated pig, the ratio was relatively unchanged (1.56 ± 0.03 and 1.43 ± 0.06 at 50 and 0.5 µM, respectively). In a human liver microsomal sample, the ratio increased from 0.13 ± 0.01 to 0.32 ± 0.05, similar to results from other laboratories.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Several investigators have demonstrated the induction of P-450 isozymes in cultured hepatocytes, e.g., induction of CYP1A isozymes by -naphthoflavone and CYP2B1/2 by phenobarbital in cultured rat hepatocytes (Wortelboer et al., 1987; Waxman et al., 1990). Li et al. (1995b) demonstrated the induction of CYP3A4 by rifampin in cultured human hepatocytes. In addition to the examples reported above, the inducibility of P-450 isozymes in cultured hepatocytes from several different species has been studied (Daujat et al., 1991; Donato et al., 1994, 1995). Multiplicity in CYP3A isoforms has been detected in several species. In humans, the genes for three CYP3A isoforms, CYP3A4, CYP3A5, and CYP3A7 have been cloned (Gonzalez et al., 1985; Aoyama et al., 1989; Komori et al., 1989). In addition, multiplicity of CYP3A isoforms has also been reported in several other species (Nelson et al., 1996 and references therein). Some of the CYP3A isoforms are inducible, some are constitutively expressed, and others are fetal forms. As shown in Fig. 4, four putative immunoreactive proteins were detected in pig liver microsomes with an anti-human CYP3A4 antibody. The two most intensely staining bands were also inducible with rifampin.

Several investigators have demonstrated that MDZ is a specific substrate probe for CYP3A isozymes (Fabre et al., 1988a; Kronbach et al., 1989; Gorski et al., 1994; Ghosal et al., 1996). Rifampin and dexamethasone were selected as potential inducers of CYP3A-mediated metabolism of MDZ in cultured porcine hepatocytes based on species-specific response to these inducers (Kocarek et al., 1995). A 6-fold induction in the formation rate of 1'-OH MDZ and a 9-fold induction in the formation rate of 4-OH MDZ in rifampin-treated porcine hepatocyte cultures over control cultures were observed. No increase was found in the formation rates in dexamethasone-treated cultures over control cultures. Thus MDZ metabolism responds to rifampin and dexamethasone pretreatment analogous to the response of rabbit and human CYP3A-mediated metabolism, but is opposite to that of rats (Kocarek et al., 1995).

Fabre et al. (1988b) demonstrated that the major metabolite of MDZ produced upon incubation of MDZ with freshly isolated human hepatocytes was 1'-OH MDZ. In addition, 4-OH MDZ and 1',4-diOH MDZ were also detected and the rate of formation of 1'-OH MDZ was roughly 15-fold greater than the rate of formation of 4-OH MDZ in media. Our observation that MDZ is metabolized to 1'-OH MDZ, 4-OH MDZ, and possibly 1',4-diOH MDZ in pigs is consistent with these results. However, the rate of formation of 1'-OH MDZ in cultured porcine hepatocytes was only 5-fold greater than the 4-OH MDZ formation rate. Fabre and colleagues also studied the metabolism of MDZ in microsomes isolated from rabbit livers (Fabre et al., 1988a). Studies with rabbit microsomes demonstrated that MDZ is primarily metabolized to 1'-OH MDZ. The metabolism of MDZ to 1'-OH MDZ and 4-OH MDZ was greater in microsomes prepared from the livers of rabbits that had been treated with rifampin compared with microsomes from untreated rabbits (Fabre et al., 1988a). In larger pigs, Ochs et al. (1987) reported that only 4-OH MDZ was observed in the plasma after i.v. or oral doses of 1 mg/kg MDZ (peak oral concentration = 0.6 µM). MDZ was incubated at 50, 5, and 0.5 µM with porcine microsomes to determine whether there was a concentration-dependent regioselectivity. In untreated microsomes, the ratio of 4-OH MDZ/1'-OH MDZ increased as the MDZ concentration decreased from 50 to 0.5 µM, but 1'-OH MDZ was still a significant metabolite. This ratio was relatively constant in rifampin-treated microsomes. In human liver microsomes, this ratio also decreased with decreasing concentrations (Gorski et al., 1994). Thus it would appear that pigs are capable of forming 1'-OH MDZ even at low concentrations. It is possible that 1'-OH MDZ is more rapidly glucuronidated, and thus was not observed in pig plasma as an unconjugated metabolite, or that there are strain differences between pigs. Additional examination of MDZ metabolites in urine will be necessary to determine whether 1'-OH MDZ is the major oxidative metabolite in pigs.

Recently, Ekins et al. (1998) have shown that in addition to CYP3A isoforms, MDZ could also be metabolized by human CYP2B6. However, the intrinsic clearance (V_max/K_m) reported for MDZ 1'-hydroxylation by CYP2B6 of 1.69 × 10⁴ ml/min/mg protein was 600-fold smaller than that reported by Ghosal et al. (1996) for cloned, expressed CYP3A4 (0.103 ml/min/mg protein) assuming a similar level of P-450 expression. A similar comparison based on the P-450 content can also be made. Based on a reported average expression level of 120 pmol of P-450/mg protein for lymphoblast-expressed CYP2B6 (Gentest Corporation catalog), the V_max/K_m for CYP2B6 reported by Ekins et al. (1998) was 1.41 × 10⁶ ml/min/pmol P-450. A V_max/K_m value of 1.78 × 10³ ml/min/pmol P-450 was obtained by Gibbs et al. (1999) with baculovirus insect-cell expressed CYP3A4. Furthermore in humans, the level of expression of CYP3A4 in liver is much higher than that of CYP2B6 (Shimada et al., 1994). These data suggest that in humans, CYP2B6 is unlikely to significantly contribute to MDZ 1-hydroxylation. Although it is possible that a CYP2B isoform in pigs is involved in the metabolism of MDZ, it is unlikely that the isoforms being induced are CYP2B isoforms, as evidenced by Western and Northern blotting experiments with CYP3A-selective probes.

Gorski et al. (1994) studied the regioselectivity in MDZ metabolism by human liver microsomes. Human liver microsomes containing both CYP3A4 and CYP3A5 showed a greater ratio of 1'-OH MDZ to 4-OH MDZ formation compared with microsomes containing only CYP3A4. The 1'-OH/4-OH MDZ ratio was also altered in incubations with purified CYP3A4 compared with CYP3A5. CYP3A5 is polymorphically expressed in the human liver and is reportedly not inducible by rifampin (Aoyama et al., 1989; Wrighton et al., 1989). Hence treatment of cultures with rifampin would lead to greater levels of CYP3A4 compared with CYP3A5, which would lead to greater induction of 4-hydroxylation compared with 1'-hydroxylation. Our results are consistent with these reports, suggesting that more than one isozyme from the CYP3A family may be involved in MDZ metabolism in porcine hepatocytes. This is further supported by the observation of two inducible anti-CYP3A immunoreactive proteins in rifampin-induced microsomes (Fig. 3).

Western blot analysis revealed a greater than 10-fold induction in the levels of a CYP3A immunoreactive protein (band 2, ~52 kDa) and a 6-fold increase in another CYP3A immunoreactive protein (band 3, 51.5 kDa) in microsomes from rifampin-induced cultures compared with uninduced microsomes. Microsomes from rifampin-treated pigs contained approximately a 4-fold increase in band 2 and a 6-fold increase in band 3 compared with microsomes from uninduced pigs. Two bands (bands 1 and 4) with low staining intensity in pig liver microsomes were not observed in either control or induced microsomes in culture. These results suggest that there is either differential stability or differential expression of CYP3A isoforms in porcine hepatocytes after rifampin treatment. Several other groups have reported that CYP3A expression is lost after several days in both rat and human hepatocyte cultures (Daujat et al., 1991; Kocarek et al., 1992). The antibody used in the Western blotting experiments was raised against human CYP3A4 and was purified such that immunoreactivity was observed only with CYP3A isoforms in humans (K. Thummel, personal communication). Thus the two major bands that respond to rifampin induction in both pigs and in culture are most likely CYP3A proteins. Bands 1 and 4 were diffuse and faint in pig liver microsomes. It is possible that these bands may be due to nonspecific cross-reactivity with other porcine P-450s (because the mw range is in the reported range), or that these are other nonP-450 pig liver proteins that might cross-react with the secondary antibody.

Witkamp et al. (1995) reported on the effect of tiamulin (an inducer of CYP3A isoforms) in cultured porcine hepatocytes and porcine liver microsomes. The presence of tiamulin led to an increase in the levels of a CYP3A immunoreactive protein in cultured hepatocytes, as determined by immunoblotting experiments with a monoclonal antibody raised against rat CYP3A1/2. In contrast to our results, Witkamp et al. (1995) only observed the presence of a single band in a mini-gel with their antibody. This discrepancy is most likely due to different cross-reactivity of the antibodies to pig CYP3A isoforms. It should be noted that our immunoblotting experiments were performed with a polyclonal antibody that is known to react with human CYP3A4, CYP3A5, and CYP3A7, whereas Witkamp and colleagues (1995) used a monoclonal antibody raised against rat CYP3A1/2.

van den Broek et al. (1981) studied the effect of the administration of rifampin on the in vivo metabolism of antipyrine and hexobarbital in pigs. Treatment of pigs with rifampin led to an increase in the metabolism of hexobarbital and antipyrine. Antipyrine is metabolized to 4-hydroxyantipyrine by several isozymes of the P-450 family, including CYP3A (Sharer and Wrighton, 1996). In addition, van den Broek et al. (1984) measured the levels of microsomal proteins from rifampin-treated pigs and untreated pigs by gel electrophoresis-Coomassie Blue staining. In microsomes from pigs treated with rifampin. there was an increase in the levels of a protein band with an approximate molecular mass of 52 kDa, and a decrease in the levels of a protein with an approximate molecular mass of 55 kDa, compared with untreated pigs. Our results indicate a similar profile of expression of CYP3A isozymes in cultured porcine hepatocytes. We observed an increase in the levels of a CYP3A isoform (~52 kDa) but no change in the levels of another CYP3A isoform (~55 kDa).

Olsen et al. (1997) have also reported the effect of treating porcine hepatocytes with various known inducers of CYP3A4 in human liver. It was observed that treating porcine hepatocytes with rifampin led to an elevation in the expression of a CYP3A immunoreactive protein. They also evaluated the effect of rifampin on the metabolism of 17--ethynylestradiol and testosterone in cultured porcine hepatocytes. There was an increase in the 2-hydroxylation of 17--ethynylestradiol and testosterone 6--hydroxylation, metabolic activities also associated with CYP3A4 in humans, after treatment with rifampin.

Northern blot analysis of mRNA isolated from rifampin-induced and control porcine hepatocyte cultures revealed an increase in the intensity of an mRNA signal (~2.4 kB) in rifampin-induced cultures compared with control cultures. Schuetz et al. (1994) demonstrated the presence of mRNA corresponding to CYP3A4, CYP3A5, and CYP3A7 genes in human livers with selective cDNA probes for the different genes. Although our results from Western blotting experiments suggested the presence of more than one isoform of CYP3A enzymes in porcine hepatocytes, we observed only one mRNA signal in the Northern blotting experiments. The cDNA probe was designed based on the sequence homology region around the conserved Cys residue in the heme binding region of CYP3A enzymes (Strömstedt et al., 1991). It is possible that multiple mRNA signals remain undetected because of the nonspecificity of the cDNA probe or that both mRNAs have a similar molecular weight, or that only one CYP3A mRNA was identified due to the high stringency conditions used in the Northern blotting experiments.

In summary, treatment of cultured porcine hepatocytes with rifampin led to an increase in the metabolism of MDZ, an increase in two immunoreactive proteins after blotting with an anti-CYP3A4 antibody, and an increase in mRNA detected with a specific CYP3A cDNA probe. Collectively, the results from the monolayer culture experiments suggest that the mechanism for CYP3A-like protein induction in porcine hepatocytes by rifampin involves regulation at the transcriptional level. The elevated mRNA levels observed in rifampin-treated hepatocytes resulted in elevated protein levels, and consequently higher rates of MDZ hydroxylation. A similar mechanism of induction involving transcriptional regulation has been proposed for CYP3A isozymes in other species (Kocarek et al., 1995). Recent evidence points to the existence of rifampin-response element in promoter region of the CYP3A4 gene, suggesting that the mechanism of induction by rifampin is through transcriptional activation (Goodwin et al., 1998). However, other mechanisms of induction such as mRNA stabilization or protein stabilization may also occur.

Based on the studies described above, porcine hepatocytes seem to behave similarly to human hepatocytes in terms of rifampin induction. Our data suggest that pigs express CYP3A-like proteins in liver. In conclusion, pig hepatocytes may offer a valuable tool to predict the metabolism of CYP3A substrates in human hepatocytes. However, additional studies are necessary to determine the homology of the porcine CYP3A isoforms with the other CYP3A isoforms from humans and other species and their substrate selectivities.