Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Inhibitors of 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMG-CoA reductase² ), also known as "statins", have achieved an important place in the arsenal of therapeutic agents available for the treatment of hypercholesterolemia, a major risk factor for the development of coronary artery disease, the leading cause of death in the United States. Five drugs of this class are currently approved for use in the United States, namely, lovastatin, simvastatin, pravastatin, fluvastatin, and atorvastatin (Food and Drug Adminstration Orange Book; http://www.fda.gov/cder/ob/default.htm). In general, these drugs have proven to be safe and effective when taken over a period of years (Davidson, 2001), and additional therapeutic uses of these drugs are under investigation (White, 1999; Puddu et al., 2001; Whitfield, 2001). However, one precaution in the use of these drugs is that most of them have been shown to interact, in one manner or another, with the cytochrome P450 system of drug-metabolizing enzymes, thereby leading to the possibility for pharmacokinetic interactions with coadministered drugs. Thus, abundant information indicates that lovastatin, simvastatin, and atorvastatin interact with human CYP3A, both as substrates and inhibitors (Prueksaritanont et al., 1997; Beaird, 2000; Cohen et al., 2000; Farmer and Torre-Amione, 2000), whereas the highest affinity interactions involving fluvastatin are with CYP2C9 (Transon et al., 1996; Beaird, 2000; Cohen et al., 2000; Farmer and Torre-Amione, 2000; Scripture and Pieper, 2001). Pharmacokinetic interactions, presumed to be based on CYP3A inhibition, have been described between simvastatin and diltiazem (Mousa et al., 2000), as well as between simvastatin or atorvastatin and the protease inhibitor nelfinavir (Hsyu et al., 2001). In contrast, CYP3A inhibitors have no effect on fluvastatin pharmacokinetics (Scripture and Pieper, 2001). Unlike the other HMG-CoA reductase inhibitors, pravastatin does not seem to interact substantially with the P450 system, either as substrate or inhibitor (Beaird, 2000; Cohen et al., 2000; Farmer and Torre-Amione, 2000), and, consistent with these findings, mibefradil, a potent CYP3A inhibitor, had no effect on pravastatin pharmacokinetics (Becquemont et al., 1999).

Also, several of the HMG-CoA reductase inhibitors have been shown to increase the expression of cytochromes P450. For example, we previously demonstrated that treatment of primary cultures of rat hepatocytes with lovastatin, simvastatin, or fluvastatin increased the levels of CYP2B, CYP3A, and CYP4A mRNA and immunoreactive protein (Kocarek and Reddy, 1996). Fluvastatin was a particularly efficacious inducer of CYP2B in the hepatocyte cultures and also increased the contents of CYP2B and CYP4A, but not CYP3A, mRNA and immunoreactive protein in the livers of treated rats (Kocarek and Reddy, 1996). In contrast, pravastatin was completely without effect on the expression of any of the P450s that were examined in the rat hepatocyte cultures (Kocarek and Reddy, 1996). Lovastatin has also been shown to induce CYP3A in primary cultured rabbit (Kocarek et al., 1995) and human (Schuetz et al., 1993) hepatocytes, as well as in HepG2 cells (Schuetz et al., 1993), whereas atorvastatin has been reported to induce CYP2C9 activity in monkey hepatocytes (Cohen et al., 2000). After the discovery that the PXR mediates the effects of many CYP3A-inducing agents, Lehmann et al. (1998) demonstrated, using a transactivation assay in CV-1 cells, that lovastatin markedly activated human PXR (also called steroid and xenobiotic receptor), and, to a lesser extent, mouse PXR. More recently, El-Sankary et al. (2001) demonstrated that lovastatin and simvastatin, but not pravastatin, activated transcription from a CYP3A4 reporter plasmid in transiently transfected HepG2 cells, and the inductions were substantially augmented by cotransfection of human glucocorticoid receptor and PXR. However, one limitation of such studies is that HMG-CoA reductase inhibitors are highly toxic to replicating cell lines, because these drugs block not only sterol production but also the prenylation reactions that are essential for cell replication.

Presently, there is no information on the effects of HMG-CoA reductase inhibitors on P450 expression in primary cultured human hepatocytes, other than the aforementioned effects of lovastatin on CYP3A. Also, no studies in a human system have been performed that attempt to evaluate whether the effects of these agents on P450 expression are the result of cholesterol biosynthesis inhibition, or whether the effects are more likely the result of direct drug-mediated activation of the PXR or other xenobiotic-sensing receptor. Thus, the purpose of this study was first, to evaluate the effects of members of this important class of therapeutic agents on P450 expression in primary cultured human hepatocytes, and second, to take advantage of known pharmacological differences among this family of drugs, by using them to query the mechanisms by which HMG-CoA reductase inhibitors regulate P450 expression.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials. Lovastatin, simvastatin, pravastatin, and fluvastatin were obtained from the sources described previously (Kocarek and Reddy, 1996). Atorvastatin and inactive enantiomer were gifts from Pfizer Global Research and Development (Ann Arbor, MI). Matrigel was purchased from Collaborative Research (Bedford, MA). Cell culture supplies and protein molecular weight markers were purchased from Invitrogen (Carlsbad, CA). Precast polyacrylamide gels were purchased from Bio-Rad (Hercules, CA). ECL Western Blotting Detection Reagents were purchased from Amersham Biosciences (Piscataway, NJ). Other supplies were obtained from the sources described previously (Kocarek and Reddy, 1996).

Isolation and Primary Culture of Human Hepatocytes. Three high-quality human livers that were judged to be unsuitable for transplantation were obtained from the Transplant Society of Michigan, and hepatocytes were prepared under a protocol approved by the Wayne State University Human Investigation Committee, as described recently (Duanmu et al., 2002). After the final wash, hepatocytes were suspended in Williams' Medium E containing 0.25 U/ml insulin, 10⁷ M triamcinolone acetonide, penicillin, and streptomycin (defined as standard Williams' Medium E), and yield and viability were estimated by counting trypan blue-stained samples, using a hemocytometer. The mean yield was 1.5 × 10⁹ viable hepatocytes, with an average viability at isolation of 78%. The hepatocytes were diluted into standard Williams' Medium E containing triamcinolone acetonide and 10% fetal bovine serum and were plated at 3 million viable cells/dish onto 60-mm dishes that were precoated with 1.5 mg Matrigel, as described previously (Kocarek and Reddy, 1996). After 3 to 10 h, the medium was replaced with standard Williams' Medium E, but lacking serum. Forty-eight to 72 h after plating, cultures were treated with drugs (3 to 5 dishes/treatment group), as described in the legend to Fig. 1. Twenty-four hours after treatment, hepatocytes were harvested for measurement of P450 and HMG-CoA reductase mRNA levels by Northern blot hybridization.

View larger version (60K): [in this window] [in a new window]   Fig. 1.   Effects of HMG-CoA reductase inhibitor treatments on CYP2B6 and CYP3A mRNA levels in primary cultures of human hepatocytes. Freshly isolated hepatocytes were prepared from three different human livers and placed into primary culture, as described under Materials and Methods. Approximately 48 h after plating, the cultures were incubated with Williams' Medium E alone (UT) or containing 0.1% DMSO (vehicle control), 105 M -naphthoflavone (BNF, in DMSO), 104 M phenobarbital (PB, in water), 105 M dexamethasone (Dex, in DMSO), 104 M ciprofibrate (Cipro, in DMSO), 3 × 105 M lovastatin (Lova, in aqueous vehicle), 3 × 105 M simvastatin (Simva, in aqueous vehicle), 3 × 105 M pravastatin (Prava, in water), 3 × 105 M fluvastatin (Fluva, in water), or 3 × 105 M of the active (Atorva) or inactive (3S,5S-Atorva) enantiomer of atorvastatin (in DMSO). Twenty-four hours after treatment, hepatocytes were harvested for the preparation of total RNA, followed by analysis of CYP2B6, CYP3A, and HMG-CoA reductase (HMGRed) mRNA levels by Northern blot hybridization. After initial hybridization of membranes with the CYP2B6, CYP3A or HMG-CoA reductase cDNA probe, blots were rehybridized with a cDNA to 7S RNA, to permit normalization for loading and transfer among samples. Left, autoradiographs of Northern blots from each set of cultured human hepatocytes. Right, autoradiographic data were quantified by scanning laser densitometry and normalized across experiments for statistical analysis as described under Materials and Methods.*, p < 0.05 versus the appropriate negative control group (i.e., UT or DMSO). , p <0.05 versus the Atorva(+) group.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Effects of HMG-CoA reductase inhibitor treatments on CYP2B6 and CYP3A immunoreactive protein levels in primary cultures of human hepatocytes. Flasks containing human hepatocytes attached to a collagen substratum, prepared from a single donor, were maintained in primary culture, as described under Materials and Methods. Approximately 48 h after plating, the cultures were overlaid with Matrigel, and 72 h after plating were incubated with Williams' Medium E alone (UT), or containing 0.1% DMSO (vehicle control), 105 M -naphthoflavone (BNF, in DMSO), 104 M phenobarbital (PB, in water), 105 M dexamethasone (Dex, in DMSO), 104 M ciprofibrate (Cipro, in DMSO), 3 × 105 M lovastatin (Lova, in aqueous vehicle), 3 × 105 M simvastatin (Simva, in aqueous vehicle), 3 × 105 M pravastatin (Prava, in water), 3 × 105 M fluvastatin (Fluva, in water), or 3 × 105 M of the active (Atorva) or inactive (3S,5S-Atorva) enantiomer of atorvastatin (in DMSO). Twenty-four hours after treatment, hepatocytes were harvested for the preparation of microsomes, followed by analysis of CYP2B6 and CYP3A immunoreactive protein levels by Western blot hybridization and enhanced chemiluminescent detection. The standards consist of microsomes containing 0.5 pmol of expressed recombinant P450s.

Primary Culture of Rat Hepatocytes. Hepatocytes were isolated from the livers of adult male Sprague-Dawley rats (approximately 300 g) under a protocol approved by the Wayne State University Animal Investigation Committee, as described previously (Kocarek and Reddy, 1996). After isolation, 3 million viable hepatocytes were plated onto 60-mm Matrigel-coated dishes and maintained in standard Williams' Medium E, as described above. Forty-eight hours after plating, cultures were treated with drugs for 24 h (3 dishes/treatment group), as described in the legend to Fig. 3.

View larger version (58K): [in this window] [in a new window]   Fig. 3.   Effects of atorvastatin treatments on P450 mRNA levels in primary cultured rat hepatocytes. Freshly isolated hepatocytes were prepared from two different rat livers and placed into primary culture, as described under Materials and Methods. Forty-eight hours after plating, the cultures were incubated with Williams' Medium E alone (UT), or containing 0.1% DMSO (vehicle control), 105 M -naphthoflavone (BNF, in DMSO), 104 M phenobarbital (PB, in water), 105 M dexamethasone (Dex, in DMSO), 104 M ciprofibrate (Cipro, in DMSO), 3 × 105 M fluvastatin (Fluva, in water), or 3 × 105 M of the active (Atorva) or inactive (3S, 5S-Atorva) enantiomer of atorvastatin (in DMSO). Twenty-four hours after treatment, hepatocytes were harvested for the preparation of total RNA, followed by analysis of CYP1A1, CYP2B, CYP3A, CYP4A, and HMG-CoA reductase (HMGRed) mRNA levels by Northern blot hybridization. After initial hybridization of membranes with cDNA probe, blots were rehybridized with a cDNA to 7S RNA, to permit normalization for loading and transfer among samples. Left, autoradiographs of Northern blots from one rat hepatocyte culture experiment. Right, autoradiographic data from each experiment were quantified by scanning laser densitometry and normalized, as described under Materials and Methods, except that the PB, Dex, Cipro, and Fluva groups were used as the positive control groups for CYP2B, CYP3A, CYP4A, and HMG-CoA reductase, respectively. Each column represents the mean relative mRNA value ± range.

Northern Blot Analysis. The three to five dishes of hepatocytes constituting each treatment group were pooled for the preparation of total RNA, as described previously (Kocarek and Reddy, 1996). Ten-microgram samples of the pooled RNAs were resolved on denaturing agarose gels and analyzed by Northern blot hybridization, as described previously (Kocarek and Reddy, 1996). cDNA probes to CYP1A1, CYP2B6, and CYP3A7 were gifts from Drs. John J. Reiners, Jr. (Wayne State University, Detroit, MI), Frank Gonzalez (National Cancer Institute, Bethesda, MD), and Erin Schuetz (St. Jude Children's Research Hospital, Memphis, TN), respectively. A cDNA corresponding to base pairs 580 to 2613 of CYP4A11 (GenBank GI 13435387) was prepared by reverse transcriptase-polymerase chain reaction amplification, using human hepatic total RNA as template. A cDNA probe to human HMG-CoA reductase (pHRed-102) was purchased from the American Type Culture Collection (Manassas, VA). Other cDNA probes were obtained from the sources described previously (Kocarek and Reddy, 1996). After hybridization with the P450 or HMG-CoA reductase cDNA probes, hybridizable bands were identified by autoradiography, and their intensities were estimated by scanning laser densitometry (Molecular Dynamics, Sunnyvale, CA). Radiolabeled probes were then removed from the filters by incubation in 1% SDS at 90°C, and blots were rehybridized with 7S cDNA, to control for RNA loading and transfer. To permit statistical analysis of the human hepatocyte Northern blot data, each data point (i.e., each P450 or HMG-CoA reductase band) on a blot was first normalized to the corresponding amount of 7S RNA detected in that sample. Second, to normalize data across the three experiments, each data point was then calculated as a percentage of the amount of RNA that was present in an appropriate "positive control group" (i.e., the phenobarbital-treated group for CYP2B6 and CYP3A4, and the pravastatin-treated group for HMG-CoA reductase). Although this calculation fixed the amount of mRNA present in the positive control group at 100%, it permitted calculation of a mean mRNA level ± S.D. for each of the other treatment groups. These values were then analyzed by one-way analysis of variance followed by the Newman-Keuls multiple comparison test, with p < 0.05 considered to be statistically significant.

Western Blot Analysis. The two flasks of hepatocytes constituting each treatment group were pooled for the preparation of microsomes, as described previously (Kocarek and Reddy, 1996). Either 10 µg (for measurement of CYP2B6 levels) or 1 µg (for measurement of CYP3A levels) of microsomal protein was resolved by SDS-PAGE (10% acrylamide) and electrophoretically transferred to nitrocellulose filters, using a mini-Protean electrophoresis unit (Bio-Rad), as described previously (Kocarek and Reddy, 1996). A polyclonal antibody directed against a CYP2B6 peptide (WB-2B6-PEP, catalog number 458226) and a monoclonal antibody directed against human CYP3A proteins (WB-MAB-3A, catalog number 458254) were purchased from Gentest (Woburn, MA), as were protein standards (i.e., human CYP2B6 microsomes and human CYP3A4+P450 reductase microsomes). Membranes were blocked and developed with antibodies, as recommended by Gentest, except that the primary antibody concentrations were reduced to 1:6,000 for WB-2B6-PEP and 1:20,000 for WB-MAB-3A, and incubations were conducted overnight at 4°C. Horseradish peroxidase-conjugated secondary antibodies, obtained from Jackson Immunoresearch Laboratories (West Grove, PA), consisted of goat anti-rabbit (for CYP2B6, used at 1:10,000) and goat anti-mouse (for CYP3A, used at 1:20,000). After hybridization with antibodies, immunoreactive bands were identified by enhanced chemiluminescence, and intensities were estimated by scanning laser densitometry.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

We previously reported that treatment of primary cultured rat hepatocytes with lovastatin, simvastatin, or fluvastatin resulted in induction of CYP2B, CYP3A, and CYP4A mRNAs and immunoreactive proteins in primary cultured rat hepatocytes (Kocarek and Reddy, 1996). Fluvastatin was a particularly effective CYP2B inducer in the rat hepatocyte cultures (Kocarek and Reddy, 1996, 1998) and also increased CYP2B and CYP4A mRNA and protein content in the livers of treated rats (Kocarek and Reddy, 1996). In contrast, pravastatin was completely without effect on P450 expression in the rat hepatocyte cultures, despite effectively up-regulating HMG-CoA reductase mRNA levels, thereby indicating that the drug inhibited cellular cholesterol biosynthesis (Kocarek and Reddy, 1996). To determine whether these effects on P450 expression are conserved in human hepatocytes, we have examined the effects of treatments with these agents on P450 expression in primary cultures of human hepatocytes. To use the human hepatocytes efficiently, we designed our experiments based on the information we had gained from our previous rat studies. Thus, because the maximally effective concentration of the HMG-CoA reductase inhibitors in rat hepatocyte cultures was 3 × 10⁵ M, we used this concentration in our experiments. Also, because the effects on P450 expression in the rat cultures were evident at both the mRNA and protein levels, and because one of our primary goals is to understand the mechanism(s) by which these drugs regulate P450 gene expression, we focused primarily on the effects of treatments on mRNA levels in the present study.

We also wished to extend our previous study to examine the effect of an additional HMG-CoA reductase inhibitor, atorvastatin. This synthetic HMG-CoA reductase inhibitor was approved for the treatment of hypercholesterolemia in 1997 (Chong and Seeger, 1997) and has since achieved widespread use, ranking as one of the most prescribed drugs in the United States in 2000 (the Internet drug index: http://www.rxlist.com/top200.htm, site viewed on February 25, 2002). An additional advantage of atorvastatin, for experimental purposes, is that the drug is a pure enantiomer (+), and an inactive enantiomer () of atorvastatin (i.e., that does not inhibit HMG-CoA reductase) was available for use. This provided us with an opportunity to determine, using classical pharmacological criteria, whether effects on P450 expression are linked to inhibition of HMG-CoA reductase activity. Although enantiomers of fluvastatin have also been described (Transon et al., 1996), we were unable to obtain these for our studies. To consider the possibility for substantial interindividual differences in drug effects, we repeated the study in hepatocyte cultures prepared from three different human livers, and results from all three hepatocyte culture experiments are shown (Fig. 1, left). Despite the concern that differences in liver quality, or pharmacogenetic differences among individuals, might lead to marked differences in outcome among experiments, highly consistent effects were seen across the three human hepatocyte culture experiments. Thus, in each human hepatocyte preparation, treatment with phenobarbital, which served as a positive control for both CYP2B6 and CYP3A³ mRNA induction, consistently elevated the levels of these mRNAs. These results support other recent observations that CYP2B6 is a highly inducible P450 (Strom et al., 1996; Chang et al., 1997; Gervot et al., 1999; Pascussi et al., 2000; Gerbal-Chaloin et al., 2001; Goodwin et al., 2001). Also, in each preparation, treatment with lovastatin, simvastatin, fluvastatin, or atorvastatin, but not pravastatin, increased the content of CYP2B6 and CYP3A mRNAs. All of the increases (3.8- to 9.2-fold for CYP2B6; 24- to 36-fold for CYP3A) were statistically significant, relative to vehicle-treated controls, with the sole exception that the effect of simvastatin treatment on CYP2B6 mRNA content, which was still 4.5-fold, did not achieve statistical significance (Fig. 1, right). Treatment with each of the active HMG-CoA reductase inhibitors produced the expected compensatory increases in HMG-CoA reductase mRNA levels, which are secondary to diminished sterol synthesis and content, followed by activation of the SREBP family of transcription factors (Edwards et al., 2000). In support of this assessment, treatment with the inactive atorvastatin enantiomer did not increase HMG-CoA reductase mRNA levels in the hepatocyte cultures. Of note, however, treatment with either atorvastatin or its inactive enantiomer produced equivalent increases in the amounts of both CYP2B6 and CYP3A mRNAs.

As an additional observation, treatment with -naphthoflavone consistently and significantly increased the amount of CYP2B6 mRNA in the human hepatocyte cultures (Fig. 1). To our knowledge, such an effect of -naphthoflavone, which does not occur for CYP2B mRNA in cultured rat hepatocytes (Fig. 3), has not been reported previously. Finally, none of the HMG-CoA reductase inhibitors produced any consistent effect on CYP1A1 or CYP4A11 mRNA levels in the primary cultured human hepatocytes (data not shown).

To determine whether effects of HMG-CoA reductase inhibitors on P450 expression also occurred at the protein level, an additional preparation of human hepatocytes was treated with the drugs, and amounts of CYP2B6 and CYP3A immunoreactive protein levels were estimated by Western blot hybridization (Fig. 2). Although Western blots developed with the WB-2B6-PEP antibody exhibited some nonspecific cross-reactivity, increases relative to untreated control in the amount of immunoreactive protein comigrating with the CYP2B6 standard were clearly evident in cultures treated with lovastatin, simvastatin, or fluvastatin, whereas no increase was observed in pravastatin-treated cultures. Also, as noted on the Northern blots, -naphthoflavone treatment produced a definite increase in the amount of CYP2B6 protein. Results obtained with the WB-MAB-3A antibody indicated that treatment with lovastatin, simvastatin, fluvastatin, and the atorvastatin isomers all produced marked increases in the amount of CYP3A immunoreactive protein (ranging from ~6- to 20-fold), whereas pravastatin treatment again had no effect. These findings indicated that the effects of the HMG-CoA reductase inhibitors on P450 expression that were observed at the mRNA level were generally maintained at the protein level.

In our previous report (Kocarek and Reddy, 1996), we did not determine the effects of atorvastatin on P450 expression in primary cultured rat hepatocytes. We therefore examined the effects of treatment with this agent, or its inactive enantiomer, on P450 mRNA content in rat hepatocyte cultures, with fluvastatin treatment serving as a control for comparison (Fig. 3). Although fluvastatin treatment produced the previously reported strong increases in CYP2B and CYP4A mRNA levels, with a modest increase in CYP3A mRNA content (Kocarek and Reddy, 1996), treatment with either atorvastatin or its inactive enantiomer had no effect on either CYP2B or CYP4A mRNA levels, but rather evoked a marked increase in CYP3A mRNA levels, which approached 50% of the amount induced by dexamethasone treatment. Again, as expected, only treatment with fluvastatin or active atorvastatin, but not the inactive atorvastatin enantiomer, caused a compensatory increase in HMG-CoA reductase mRNA content. Thus, the profile of P450 induction that was generated by atorvastatin treatment of rat hepatocytes differed from that produced by any of the other HMG-CoA reductase inhibitors that we have examined (Fig. 3; Kocarek and Reddy, 1996).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

A major goal of this study was to obtain insight into whether the effects of HMG-CoA reductase inhibitors on P450 gene expression are mechanistically linked to HMG-CoA reductase inhibition. Thus, even though lovastatin was previously shown to activate the human PXR (Lehmann et al., 1998), it remained possible that this effect was mediated indirectly, as a result of cellular responses that were triggered by HMG-CoA reductase inhibition. In this regard, it is now well established that there is substantial interplay among the pathways involved in maintaining cellular lipid homeostasis. As two examples, Kim et al. (1998) reported that expression of SREBP1c in 3T3-L1 preadipocytes resulted in the production of an endogenous ligand for peroxisome proliferator-activated receptor . Also, several groups have reported that SREBP1c expression is dependent on the ongoing activation of the liver X receptor by oxysterols (Repa et al., 2000; DeBose-Boyd et al., 2001; Yoshikawa et al., 2001). Because the PXR, which is also activated by secondary bile acids such as lithocholic acid (Staudinger et al., 2001b; Xie et al., 2001), has now been demonstrated to play a role in regulating hepatic sterol/bile acid homeostasis (Staudinger et al., 2001a), it seems possible that disruption of cellular lipid metabolism by HMG-CoA reductase blockade could trigger an endogenous mechanism that leads to the activation of PXR. Our results, in which both atorvastatin and its inactive enantiomer increased CYP2B and CYP3A mRNA content with equal ability, and in which pravastatin effectively blocked cholesterol biosynthesis but did not elevate P450 mRNA levels, strongly indicate that ability of a drug to inhibit HMG-CoA reductase activity does not predict its ability to induce CYP2B or CYP3A in cultured human hepatocytes.

Based on our findings, we conclude the following about the degree of conservation of the effects of HMG-CoA reductase inhibitor treatments on P450 expression between rat and human. First, fluvastatin is a highly effective CYP2B inducer in both rat and human hepatocytes, with lovastatin and simvastatin being somewhat less effective, especially in rat. In contrast, fluvastatin is no more effective as a CYP3A inducer than any of the other HMG-CoA reductase inhibitors tested in the hepatocytes of either species. Second, pravastatin treatment does not increase CYP2B or CYP3A mRNA or immunoreactive protein content in either rat or human hepatocytes, although it clearly gains entry into the cell and inhibits cholesterol biosynthesis in both cases. In fact, pravastatin treatment regularly produced the largest increases in HMG-CoA reductase mRNA levels that were observed in the human hepatocyte cultures, causing us to use this treatment as our standard for data normalization. Overall, therefore, pravastatin exhibits very little interaction with the P450 system, because previous studies have demonstrated that pravastatin undergoes limited P450-mediated metabolism and does not substantially inhibit the activity of any human P450 (Beaird, 2000; Cohen et al., 2000; Farmer and Torre-Amione, 2000). Also, our findings in primary cultured human hepatocytes are in close agreement with the recently reported abilities of lovastatin and simvastatin, but not pravastatin, to activate transcription from a CYP3A4 reporter plasmid in transiently transfected HepG2 cells (El-Sankary et al., 2001).

A major difference in the effects of the drugs across species was seen for atorvastatin, which increased the amount of only CYP3A mRNA in the cultured rat hepatocytes, but increased both CYP2B6 and CYP3A mRNAs in the human hepatocyte cultures. Based on current mechanistic models of CYP2B and CYP3A induction, in which PXR is the major nuclear receptor mediating the effects of xenobiotics on CYP3A expression (Lehmann et al., 1998; Jones et al., 2000), whereas CAR is the principal receptor governing the effects of phenobarbital and other agents on CYP2B expression (Wei et al., 2000), one may speculate that lovastatin, simvastatin, and fluvastatin are able to activate both CAR and PXR in rat hepatocytes, whereas atorvastatin and its inactive enantiomer activate only PXR. In contrast, in human hepatocytes, atorvastatin may be able to activate both CAR and PXR. Alternatively, PXR has been shown to play a major role in the xenobiotic-inducible regulation of CYP2B6 (Goodwin et al., 2001). Thus, PXR, rather than CAR, may be the dominant transcription factor responsible for mediating HMG-CoA reductase inhibitor-inducible expression of CYP2B6. These possibilities are currently under investigation.