If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Clinical Pharmacology, Georg-August University Goettingen, Robert-Koch-Strasse 40, D-37085 Goettingen, GermanyDepartment of Pharmacology, Johannes Gutenberg University Mainz, Obere Zahlbacher Strasse 67, D-55101 Mainz, Germany
* This work was supported by Grants Bu 1249/1-3 and WO505/2-1 from the Deutsche Forschungsgemeinschaft (Germany), by the Robert Bosch Foundation (Germany), and by Grants AA08990 and GM49511 from the National Institutes of Health (to J. R.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Induction of cytochrome P450 3A (CYP3A) by xenobiotics may lead to clinically relevant drug interactions. In contrast with other CYP3A family members, studies on the inducibility of CYP3A5 indicate conflicting results. We report the induction of CYP3A5 mRNA in 13 of 16 hepatocyte preparations exposed to rifampin. Furthermore, induction of CYP3A5 mRNA was observed in intestinal biopsies in three of eight probands following exposure to the antibiotic. The highest absolute levels of CYP3A5 transcripts were found following rifampin treatment in hepatocytes and intestines from carriers of CYP3A5*1 alleles. Elucidation of the mechanism involved in CYP3A5 induction revealed that constitutively activated receptor (CAR) and pregnane X receptor (PXR) transactivated the CYP3A5 promoter (–688 to +49) and that the transactivation was dependent on an everted repeat separated by 6 bp (ER6-dependent). Treatment with the prototypical PXR ligand rifampin led to a 2-fold induction of the CYP3A5 promoter activity. In agreement with these observations, PXR and CAR bound specifically to the ER6 motif. Hepatic expression of PXR correlated with that of CYP3A5 mRNA levels in a bank of liver samples. Taken together, studies here revealed the presence of a functional ER6 motif in the CYP3A5 promoter located –100 bp upstream from the transcription start site, suggesting that CYP3A5 is inducible by mechanisms similar to those involved in CYP3A4 induction. Enhanced expression of CYP3A5 caused by exposure to inducers may phenocopy the effects of the high expression allele CYP3A5*1. In this manner, induction of CYP3A5 may contribute to the overall importance of this P450 in drug metabolism and drug interactions.
Members of the human cytochrome 3A subfamily of P450 monooxygenases (CYP3A)
The abbreviations used are: CYP, cytochrome(s) P450; CAR, constitutively activated receptor; PXR, pregnane X receptor; ER6, everted repeat separated by 6 bp; RXRα, retinoid X receptor-α; XREM, xenobiotic-responsive enhancer module; HNF4, hepatocyte nuclear factor-4; CITCO, 6-(4-chloropheny-l)imidazo[2,1-b][1,3]thiazole-5-carbaldehyde O-(3,4-dichlorobenzyl)oxime.
1The abbreviations used are: CYP, cytochrome(s) P450; CAR, constitutively activated receptor; PXR, pregnane X receptor; ER6, everted repeat separated by 6 bp; RXRα, retinoid X receptor-α; XREM, xenobiotic-responsive enhancer module; HNF4, hepatocyte nuclear factor-4; CITCO, 6-(4-chloropheny-l)imidazo[2,1-b][1,3]thiazole-5-carbaldehyde O-(3,4-dichlorobenzyl)oxime.
play an important role in drug metabolism because of their hepatic and intestinal abundance and because of the large substrate spectrum. CYP3A enzymes account for 50% of the total P450 in an average human liver, and these enzymes participate in the metabolism of up to 50% of the drugs currently in use (
). Not surprisingly, CYP3A are frequently involved in drug-drug interactions. CYP3A inhibition by drugs such as ketoconazole or erythromycin is considered the most relevant mechanism. For example, the resulting increased levels of the co-administered nonsedating antihistamine terfenadine in the presence of ketoconazole inhibited a cardiac potassium channel, leading to serious and, in some cases, fatal cardiac arrhythmias (
). CYP3A induction is frequently considered clinically less important than inhibition because it is expected to reduce the efficacy of co-administered CYP3A substrate drugs rather than to affect their safety. However, CYP3A inducers such as rifampin and rifabutin can reduce plasma concentrations of certain drugs up to 40-fold, effectively abolishing their efficacy (
For most compounds tested, induction is mediated by the nuclear receptors PXR and/or CAR. The transcriptional activation by PXR is triggered by the binding of an inducer to the receptor and involves specific binding sites in the respective CYP3A promoter. There are four CYP3A genes in humans, although only two protein products (CYP3A4 and CYP3A5) have been identified unequivocally in adult liver in vivo. Xenobiotic-mediated increases in expression of CYP3A4 mRNA and protein and CYP3A7 and CYP3A43 mRNA have been reported (
) reported a similar result for both primary human hepatocytes and for the hepatocellular carcinoma-derived cell line TONG/HCC. On the other hand, two other more recent studies showed that CYP3A5 was inducible in human hepatocytes by rifampin and phenobarbital (
). Altogether, the data on the inducibility of CYP3A5 by typical CYP3A inducers are still conflicting.
A clarification of the inducibility of CYP3A5 is important for the assessment of the overall importance of CYP3A5 in hepatic and intestinal drug metabolism. The dominant position of CYP3A4 within the CYP3A family has been challenged recently in favor of CYP3A5, especially in individuals with increased expression of CYP3A5 because of the inheritance of CYP3A5*1 alleles (
). Inducibility of CYP3A5 could, at least in part, explain these conflicting assessments. In the following report, we present the results of our analysis of CYP3A5 induction, which involved ex vivo, in vivo, and in vitro experiments.
Hepatocytes—The two sources and preparations of primary human hepatocytes were as described previously (
) for HH livers. Drugs to be tested were dissolved in Me2SO (final concentration of Me2SO in the cell culture medium, 0.1%). Following the appropriate incubation time, cells were lysed as described, and total RNA was isolated using RNeasy kits (Qiagen, Hilden, Germany).
Intestinal Biopsies—Biopsy specimens of the duodenal mucosa were obtained by esophagogastroduodenoscopy and snap-frozen in liquid nitrogen for RNA analysis. The details of the study have been described elsewhere (
). After 9 days of oral treatment with 600 mg/day rifampin, a second biopsy was obtained and treated as described above.
Liver Samples—Liver samples were collected during surgical interventions conducted at the Department of Surgery, University Medical Center Charité, Humboldt University in Berlin, Germany. The donors were white Europeans. The liver samples included nontumorous tissue surrounding primary liver tumors and metastases of various tumors or liver material surgically removed for other reasons. Medical histories were evaluated in respect to treatment with known CYP3A inducers. The preparations of total RNA for these liver samples have been described in a previous study (
Ethics—All studies were approved by the local ethics boards, and written consent was obtained from all patients and probands.
TaqMan Assays and Genotyping—Total RNA was isolated from lysed hepatocytes and intestinal biopsies using RNeasy kits (Qiagen). The concentration of total RNA was determined spectrophotometrically, and the quality was evaluated by inspection of the integrity of 28 S and 18 S rRNA bands on agarose gels. cDNA was synthesized from 1 μg of total RNA using random hexamer primers following a procedure described previously (
). An increase in transcript numbers by a factor of 1.5 or more was considered induction. The genotype of the CYP3A5*1 allele, predictive of increased CYP3A5 expression, was determined as described previously (
). The proximal CYP3A5 promoter (–688 to +49, with respect to the transcription start site) was amplified out of human genomic DNA by PCR with the forward oligonucleotide 5′-ACA GGT ACC ATC ACC ACA GAG TCA GAG GGG ATG-3′ and the reverse primer 5′-ACA AGA TCT GCT GTT TGC TGG GCT GTT TGC CTG G-3′. The oligonucleotides introduce, respectively, a KpnI and a BglII restriction site on the 5′- and 3′-end of the amplified fragment. The PCR product was digested with KpnI and BglII and cloned into appropriately digested luciferase reporter gene vector pGL3-Basic (Promega). The identity of the CYP3A5 promoter within the cloned DNA fragment and the absence of mutations introduced by PCR were verified by sequencing of the resulting plasmid p3A5(–688). Mutagenesis of the ER6 motif was performed within plasmid p3A5(–688) using the QuikChange site-directed mutagenesis kit (Stratagene) and the oligonucleotide 5′-CCT CAT AGA ACA TGTTCT CAA AAG AGAACA GCA AAG GGG TGT GTG-3′ (half-sites of ER6 are shown in bold, and mutated bases are in italics and underlined). The introduction of the mutations and the absence of other undesired mutations were verified by sequencing of the resulting plasmid p3A5(–688/mER6).
A reporter gene plasmid of CYP3A4 comparable in size to p3A5(–688) was generated by unidirectional deletion of pGL3-CYP3A4(–1105) (
) with the double-stranded nested deletion kit (Amersham Biosciences). Sequencing identified a clone encompassing the CYP3A4 promoter region from –702 to +51 of the transcription start site, p3A4(–702).
Transient Transfections and Reporter Gene Assays—Transient co-transfection experiments in LS174T cells were performed as described previously (
). 6–7 h after transfection, the cells were washed with phosphate-buffered saline, supplied with fresh medium (without phenol red and supplemented with fetal calf serum that was pretreated with dextran-coated charcoal), and treated with 1 μm CITCO or 10 μm rifampin, both dissolved in Me2SO, or with an equivalent amount (final concentration, 0.1%) of Me2SO only. Reporter gene assays were performed as described previously (
). To identify statistically significant differences, one-way analysis of variance was performed with mean values of at least three independent experiments performed in triplicate using GraphPAD InStat version 3.05 for Windows 95.
Electrophoretic Mobility Shift Assays—The human PXR, CAR, and RXRα proteins were synthesized using the appropriate expression plasmids and the TnT T7 quick coupled transcription/translation system (Promega). Electrophoretic mobility shift assays were performed as described previously (
). Oligonucleotides for the CYP3A5 wild-type and mutated ER6 motifs were as follows: ER6 wild-type sense, 5′-GAT CCG AAC ATG AAC TCA AAA GAG GTC AGC AAA-3′; ER6 wild-type antisense, 5′-GAT CTT TGC TGA CCT CTT TTG AGT TCA TGT TCG-3′; ER6-mutated sense, 5′-GAT CCG AAC ATGTTC TCA AAA GAGAAC AGC AAA-3′; and ER6-mutated antisense, 5′-GAT CTT TGC TGTTCT CTT TTG AGAACA TGT TCG-3′. Half-sites of ER6 are shown in bold, and mutated bases are in italics and underlined. Retarded complexes were quantified with the BAS1800 II phosphor-storage scanner (Fuji) and AIDA software (Raytest).
Hepatocytes from a total of 18 donors (10 GH and eight HH series) were used in this study. The expression of CYP3A4 and CYP3A5 prior to culturing was investigated in five livers of the GH series (all CYP3A5*3 homozygotes). On average, CYP3A5 transcripts (3829 ± 3541/ng of total RNA (range of 812–9754)) accounted for 5.9 ± 0.06% of CYP3A4 transcripts (228,176 ± 307,507/ng of total RNA (range of 27,847–841,279)). After 24 h of culture in 5% fetal calf serum, hepatocytes were serumstarved for 48 h. Subsequently, the hepatocytes were cultured for 48 h in the same serum-free medium in the presence of an inducer or solvent only. A comparison of the CYP3A4 and CYP3A5 expression levels in solvent-treated hepatocytes with those prior to culturing revealed a differential effect of culturing on the expression of these genes. Although the expression of CYP3A5 dropped only in hepatocytes from one liver (GH4, 4-fold), the decrease in CYP3A4 expression was observed in all five livers, and it was much more pronounced (3418 ± 3552-fold (range of 479–8803-fold). Consequently, cell culturing led to a reversal of the initial expression ratio, with CYP3A4 accounting for less than 2% of CYP3A5 transcripts.
The spectra of CYP3A5 and CYP3A4 inducers were similar in that induction of either gene was observed with phenobarbital, mifepristone, clotrimazole, omeprazole, and rifampin. Although hepatocytes from subject GH9 exhibited induction of CYP3A4 and CYP3A5 by the aforementioned compounds, enhanced expression of these P450s was not observed with 2–25 μm 3-methylcholanthrene or 2–50 μm β-naphthoflavone. The extent of CYP3A4 and CYP3A5 induction among the GH hepatocyte samples was liver- and drug-specific. For example, when compared with other samples, hepatocytes from liver GH12 exhibited the greatest CYP3A5 induction following treatment with phenobarbital, mifepristone, clotrimazole, and omeprazole. However, rifampin produced a negligible affect on CYP3A5 expression in these same hepatocytes (Fig. 1). Conversely, CYP3A4 expression in hepatocytes from livers GH6 and GH8 was only weakly increased by drugs examined here. The strongest inducer of either gene in hepatocytes from GH livers was 1 mm phenobarbital (12.1 ± 9.5-fold for CYP3A5 and 262 ± 345-fold for CYP3A4; see Fig. 1 for induction values from individual livers), and the weakest was 10 μm mifepristone (3.0 ± 1.7-fold for CYP3A5 and 52 ± 99-fold for CYP3A4). Clotrimazole (100 μm) and 10 μm rifampin produced intermediate increases in CYP3A4 (126 ± 150-fold and 195 ± 263-fold, respectively) and CYP3A5 (10.9 ± 9.1-fold and 3.3 ± 1.7-fold, respectively). There was a statistically significant correlation between enhancement of CYP3A5 and CYP3A4 among the livers following treatment with 10 μm rifampin (r = 0.88, p < 0.001).
Liver GH3 was heterozygous for the CYP3A5*1 allele. Hepatocytes from this liver had the highest levels of CYP3A5 transcripts in comparison with all other GH livers both prior to (32,129 versus 5215 ± 4466) and following (88,000 versus 17,218 ± 17,248) treatment with 10 μm rifampin. We identified one additional CYP3A5*1 carrier among rifampicin-treated hepatocytes derived from eight livers of the HH series. In the HH hepatocytes, rifampin-mediated CYP3A4 induction was less than in GH hepatocytes (on average 63-fold and 195-fold, respectively), most likely because of differences in cell preparation and culturing. CYP3A5 transcripts in the HH hepatocytes were induced by rifampicin more than 1.5-fold in six of eight livers (mean, 3.7 ± 4.0-fold). The highest increase in CYP3A5 transcripts (12-fold) was observed in hepatocytes from the heterozygous CYP3A5*1 liver (HH997).
Because cell culturing may influence the -fold increase in CYP3A4 and CYP3A5 by differentially influencing basal expression levels, we investigated the induction of these genes in vivo. Biopsies of the duodenum were taken from eight healthy probands prior to and following 9 days of treatment with rifampin (600 mg/day). Genotyping revealed that proband number 6 was a heterozygous CYP3A5*1 allele carrier, whereas the other seven probands were CYP3A5*3 homozygotes. CYP3A5 transcripts were detected in all biopsies (Fig. 2), and as expected, the seven samples homozygous for CYP3A5*3 expressed the lowest levels of CYP3A5 mRNA with a mean of 1550 transcripts/ng of total RNA (range of 670–3500). In contrast, the sole CYP3A5*1 carrier (proband number 6) expressed 5750 transcripts/ng of total RNA. Following treatment with rifampin, three probands (numbers 6, 7, and 8) exhibited greater than a 2-fold enhancement of CYP3A5 mRNA expression. The highest transcript level following rifampin treatment (37,000/ng of total RNA) was again measured in the proband heterozygous for CYP3A5*1 (number 6) corresponding to a 6.4-fold increase in CYP3A5 transcript levels. The –fold induction of CYP3A5 was even higher in the CYP3A5*3-homozygous proband number 7 (9.7-fold), but the absolute transcript level was much lower (6500/ng of total RNA). Probands 6 and 7 as well as proband 2 also exhibited the highest increases in the expression of CYP3A4 mRNA (3.7-, 6.6-, and 7.9-fold, respectively) when compared with the remaining proband samples. The increase in the CYP3A4 mRNA expression in the other five probands was between 1-fold (no induction) and 4.2-fold (data not shown).
The similarities in inductive response between CYP3A4 and CYP3A5 suggested the involvement of the nuclear receptors PXR and CAR in CYP3A5 expression. These nuclear receptors mediate P450 expression by binding to specific motifs in the 5′-upstream regulatory regions of target genes. Inspection of the CYP3A5 5′-upstream region revealed the presence of a putative PXR/CAR binding site of the ER6 type around the position –100 bp from the transcription start site (
). With respect to the sequence and position, this motif is identical to the proximal ER6 motif of CYP3A4, for which binding of PXR and CAR has been demonstrated. An enhancer element homologous to the xenobiotic-responsive enhancer module (XREM) of CYP3A4 is absent from the CYP3A5 promoter (
). To determine the role of PXR in the inducibility of CYP3A4, we first investigated the binding of PXR and CAR to the ER6 motif of CYP3A5 in vitro. Fig. 3A demonstrates that PXR and CAR both bound to the ER6 motif of CYP3A5 as heterodimers with RXRα. In contrast, neither nuclear receptor bound to a mutated ER6 element (Fig. 3A). Competition gel shift experiments further demonstrated the specificity of binding; the wild-type ER6 element was able to compete for binding, and the mutated element had no effect on nuclear receptor binding to the CYP3A5 element (Fig. 3). Thus, the ER6 motif in the CYP3A5 promoter proved to be a true PXR/CAR binding site.
In subsequent experiments we determined whether the identified ER6 PXR/CAR binding site of CYP3A5 mediated enhanced activation of the CYP3A5 promoter by these nuclear receptors. LS174T cells were transiently transfected with PXR or CAR expression plasmids and luciferase reporter gene plasmids under the control of either the CYP3A5 promoter or the corresponding region of the CYP3A4 promoter takes for comparison. Activation of endogenous PXR in LS174T cells by treatment with rifampin induced CYP3A5 promoter activity 1.6-fold (p < 0.05) (Fig. 4). Co-transfection of PXR stimulated the CYP3A5 promoter 1.9-fold (p < 0.01), thereby demonstrating basal activity of PXR. Treatment of PXR-transfected cells with rifampin resulted in a further increase in CYP3A5 promoter activity by 2.1-fold (p < 0.001) (Fig. 4). Mutation of the ER6 PXR binding site in the CYP3A5 promoter abolished PXR-dependent induction by rifampin. Co-transfection of CAR resulted in a 3.1-fold (p < 0.05) activation of CYP3A5 promoter activity, which was only slightly enhanced by the human CAR-specific agonist CITCO. Again, activation by CAR was dependent on a functional ER6 motif (Fig. 4). Because we confirmed that mutation of the ER6 motif totally abolished binding of PXR and CAR (see Fig. 3), nuclear receptors PXR and CAR clearly activated the CYP3A5 promoter via the ER6 motif. For comparison, we also analyzed induction/activation of a homologous region of the CYP3A4 promoter by PXR and CAR. The CYP3A4 construct was activated 2–3-fold greater than CYP3A5 by CAR and PXR. Similarly, induction by rifampin was ∼2-fold stronger. As with CYP3A5, treatment with CITCO only weakly enhanced the activation of CYP3A4 by CAR (1.3-fold, p < 0.05).
Finally, we investigated the expression of PXR and CYP3A5 mRNA species in a bank of liver samples. The expression of CYP3A5 in these samples had been reported previously (
), and it was on average 5100 transcripts/ng of total RNA (range between 2000 and 11,000 copies) for *3/*3 livers and 41,200 transcripts (range of 21,500–62,600 transcripts/ng of RNA) in *1/*3 livers. The PXR expression was on average 11,454 transcripts/ng of total RNA and varied 10-fold within the sample set (range of 2283–23,331 transcripts/ng of total RNA). Without consideration of the CYP3A5*1 allele status, there was a weak (r = 0.22), albeit statistically significant (p < 0.05, n = 64), correlation between the PXR and CYP3A5 mRNA expression levels. The correlation coefficients increased upon inclusion of the CYP3A5*1 allele status, and they were 0.61 for *3/*3 livers (p < 0.001, n = 56) and 0.4 for *1/*3 livers (p < 0.16, n = 8) (Fig. 5). The latter correlation increased to r = 0.87 (p < 0.005, n = 7) upon exclusion of one outlier liver (Fig. 5). Altogether, these data indicated that hepatic PXR and CYP3A5 mRNA expression levels were correlated.
Previous analyses of the inducibility of hepatic CYP3A5 produced conflicting results (
) because of confusion between the real CYP3A5 5′-region and an almost identical pseudogene sequence localized ∼20 kb upstream of the CYP3A5 gene. A more recent analysis of the correct gene failed to detect rifampin-induced transcription (
). However, it should be noted that the cells used to determine functionality of the CYP3A5 promoter were not co-transfected with a PXR expression construct and that they do not express endogenous PXR (
Because of the limited amount of hepatocyte lysates, the induction was measured on the mRNA rather than on the protein level. However, CYP3A5 transcripts are a good surrogate parameter for CYP3A5 protein expression, as demonstrated in banks of liver samples (
). As demonstrated here, CYP3A5 mRNA levels were enhanced by typical CYP3A inducers known to activate PXR and/or CAR. The chemically mediated induction was clearly dependent on the presence of an ER6 binding site for PXR and CAR. In the CYP3A4 5′-upstream region, the induction by PXR or CAR can occur either by the proximal ER6 motif located at position –160 (
) suggests that HNF4 and CAR regulate the constitutive expression, whereas PXR mediates the induction by exogenous compounds.
In contrast, the 5′-regulatory region of CYP3A5 shows only a limited sequence identity to the corresponding region of CYP3A4. The ER6 motif identified here is identical to that in CYP3A4, but a distal XREM region is missing. Therefore, the induction observed in hepatocytes may be mediated exclusively by the proximal ER6 motif of CYP3A5. The mean -fold induction of CYP3A5 mRNA in hepatocytes by rifampin or mifepristone corresponded well to the enhanced activation of the CYP3A5 promoter by rifampin observed in the reporter gene assay. Thus we speculate that no further PXR/CAR binding sites exist in the 5′-upstream regulatory region of CYP3A5. However, a definitive conclusion requires further analysis of distal regions of the CYP3A5 promoter.
A major role of PXR in the hepatic CYP3A5 regulation is also supported by correlation analysis of CYP3A5 mRNA expression in relation to the CYP3A5*3 and CYP3A5*1 allele status. This analysis revealed a statistically significant correlation between CYP3A5 mRNA and that of PXR within each CYP3A5 genotype. Results presented here, together with observations demonstrating CYP3A43 (
) induction, suggest that all four CYP3A genes are inducible and that PXR is a major common regulator of this process. The co-regulation of CYP3A5 and CYP3A4 by PXR provides an explanation for the correlation between CYP3A4 and CYP3A5 mRNA expression levels demonstrated in liver samples by Lin and co-workers (
The quantitative effect of CYP3A5 induction on hepatic or intestinal CYP3A5 expression and on the overall CYP3A activity in these organs is difficult to estimate. Depending on the drug applied, the -fold induction of CYP3A5 transcripts in cultured hepatocytes is between 1.6 and 7.9% of that measured for CYP3A4. However, it should be taken into consideration that upon culturing, this is preceded by a marked (480–8800-fold) decrease in the expression of CYP3A4 transcripts. The corresponding decrease in the CYP3A5 expression is much smaller (up to 4-fold). Therefore, the much higher magnitude of CYP3A4 induction may have been caused by the initial down-regulation and may reflect the differential effect of cell culturing on these genes rather than physiological differences in their regulation. In support of this interpretation, the magnitudes of CYP3A4 and CYP3A5 induction are comparable in vivo, i.e. in intestinal biopsies from probands treated with rifampin. Taken together, these data suggest that the -fold increases in the expression levels of CYP3A4 and CYP3A5 transcripts could be comparable, at least in the intestine. Conversely, caution should be taken while extrapolating induction measurements from hepatocytes to the in vivo situation.
Our characterization of the mechanism of CYP3A5 induction may help to explain the controversy surrounding the significance of the CYP3A5 polymorphism. It has been suggested that CYP3A5 may contribute up to 50% of total hepatic CYP3A protein in individuals with increased expression of this enzyme (
) found that the CYP3A5 protein does not exceed 17% of the total hepatic CYP3A pool in CYP3A5*1 carriers and accounts for no more than 2% in an average Caucasian liver. Similarly, CYP3A5 mRNA transcripts accounted for less than 4% of the CYP3A transcript pool in two independent studies (
) on midazolam clearance in vivo. In part, these discrepancies could be caused by CYP3A5 induction that was unaccounted for in some of the liver samples investigated. Thus, the induced intestinal CYP3A5 transcript levels in some probands homozygous for CYP3A5*3 (probands 7 and 8 in Fig. 4) exceed the expression of the gene measured in the carrier of a CYP3A5*1 allele prior to rifampin treatment (proband 6). This indicates that in some CYP3A5*3 homozygotes, CYP3A inducers may phenocopy the effect of the CYP3A5*1 allele.
We are indebted to K. Abuazi de Paulus and K. Hennecke for excellent technical assistance. We thank A. K. Nuessler and P. Neuhaus (Department of Surgery, University Medical Center Charité, Humboldt University, Berlin, Germany) for collecting the liver samples.