点击显示 收起
Institut de Chimie des Substances Naturelles, Centre National de la Recherche Scientifique UPR2301, Gif-sur-Yvette, France
Abstract
The total CYP3A5 mRNA level is significantly greater in carriers of the CYP3A51 allele than in CYP3A53 homozygotes. Most of the CYP3A53 mRNA includes an intronic sequence (exon 3B) containing premature termination codons (PTCs) between exons 3 and 4. Two models were used to investigate the degradation of CYP3A5 mRNA: a CYP3A5 minigene consisting of CYP3A5 exons and introns 3 to 6 transfected into MCF7 cells, and the endogenous CYP3A5 gene expressed in HepG2 cells. The 3'-untranslated region g.31611C>T mutation has no effect on CYP3A5 mRNA decay. Splice variants containing exon 3B were more unstable than wild-type (wt) CYP3A5 mRNA. Cycloheximide prevents the recognition of PTCs by ribosomes: in transfected MCF7 and HepG2 cells, cycloheximide slowed down the degradation of exon 3B-containing splice variants, suggesting the participation of nonsense-mediated decay (NMD). When PTCs were removed from pseudoexon 3B or when UPF1 small interfering RNA was used to impair the NMD mechanism, the decay of the splice variant was reduced, confirming the involvement of NMD in the degradation of CYP3A5 splice variants. Induction could represent a source of variability for CYP3A5 expression and could modify the proportion of splice variants. The extent of CYP3A5 induction was investigated after exposure to barbiturates or steroids: CYP3A4 was markedly induced in a pediatric population compared with untreated neonates. However, no effect could be detected in either the total CYP3A5 RNA, the proportion of splice variant RNA, or the protein level. Therefore, in these carriers, induction is unlikely to switch on the phenotypic CYP3A5 expression in carriers of CYP3A53/3.
The members of the CYP3A subfamily are responsible for the metabolism of more than 50% of clinically used drugs (Eichelbaum and Burk, 2001), but the contribution of individual CYP3A isoforms is variable. The function of CYP3A43 has not yet been elucidated, but its involvement in the overall drug metabolism should be of a minor importance, as judged by its low mRNA level in different tissues (Domanski et al., 2001; Gellner et al., 2001). CYP3A7 is mostly expressed in the fetal liver and is replaced progressively by CYP3A4 after birth (Lacroix et al., 1997). CYP3A4 and CYP3A5 are likely to play the major metabolic role in adult liver. CYP3A4 is constitutively expressed in the adult liver (Westlind-Johnsson et al., 2003), whereas CYP3A5 is polymorphically expressed at an appreciable level in only a fraction of ethnic populations (Hustert et al., 2001; Kuehl et al., 2001; Lin et al., 2002).
Several authors proposed that CYP3A5 could be only a minor contributor in the overall metabolism of CYP3A substrates (Shih and Huang, 2002; Westlind-Johnsson et al., 2003; Fukuda et al., 2004). However, at least for a small number of drugs, the CYP3A5 polymorphism plays a significant role in pharmacokinetics. A clear association between tacrolimus dose requirement and the CYP3A5 genotype has been described previously (Macphee et al., 2002; Hesselink et al., 2003; Zheng et al., 2003). A dose-dependent effect was also observed with ABT-773 (Katz et al., 2004). Therefore, it is likely that the pharmacokinetics of other CYP3A substrate drugs could be dependent, at least partially, on the CYP3A5 polymorphism (Huang et al., 2004).
The low level of CYP3A5 protein expression is linked to the CYP3A53 allele. This allele displays a g.6986A>G (refSNP ID: rs776746) transition, creating a cryptic acceptor splice site in the third intron of the CYP3A5 gene (numbering according to the recommendations of the CYP Allele Nomenclature Committee, Oscarson and Ingelman-Sundberg, 2002). This cryptic site promotes the insertion of an intronic exon-like sequence (pseudoexon 3B) in the mature mRNA and could involve subsequent deletions of exon and/or insertion of other intronic sequences. These abnormally spliced messengers contain several in-frame premature termination codons (PTC), explaining the low protein level. In the past, several authors reported that CYP3A5 is not inducible (Wrighton et al., 1989), but recently, Burk et al. (2004) have demonstrated a substantial increase in CYP3A5 mRNA levels in human hepatocytes after treatment with several known CYP3A inducers. The induction of CYP3A5 could therefore constitute a non-negligible source of interindividual CYP3A5 mRNA and protein variability.
The objective of this work was to gain insight into the molecular regulation of the CYP3A5 mRNA level, which is significantly lower in CYP3A53/3 carriers than in 1 carriers. No mechanism explaining this difference has been demonstrated so far, even if several hypotheses have been proposed. First, we demonstrated that the g.31611C>T mutation (refSNP ID: rs15524) has no effect on the degradation of CYP3A5 mRNA. Then, we tested whether the lowest CYP3A5 mRNA level could result from the rapid degradation of splice variants by the nonsense-mediated mRNA decay (NMD) and the role of PTC in this mechanism. Finally, we investigated whether exposure to CYP3A inducers could modify the concentration and distribution of CYP3A5 mRNA in a panel of human livers. Because none of our adult samples was medicated, we used samples obtained from neonates or newborns treated with either barbiturates or steroids.
Materials and Methods
Human Liver Bank. Samples from white neonatal and adult subjects were obtained within the first hours after death in accordance with the recommendations of the Ethical Committee from Institut National de la Santee et de la Recherche Meedicale. Neonatal livers were obtained after informed parental consent. The treatment of newborns with barbiturates or steroids was documented, and we were allowed to classify the subjects as "induced" as reported in Cresteil et al. (1994). Adult livers came from donors for kidney transplantation exempt of chronic diseases and received no medication known for their inducible capacity of CYP3As. Samples were immediately frozen in liquid nitrogen and then stored at -80°C before analysis.
Western Blotting of CYP3A5. Human liver microsomes were prepared according to a protocol reported previously (Cresteil et al., 1979). Microsomal protein content was determined using the Uptima BCA Assay Kit (Interchim, Montluon, France). Sixty micrograms of microsomal proteins was resolved on 9% SDS-polyacrylamide gels and transferred overnight onto polyvinylidene difluoride membranes (PerkinElmer Life and Analytical Sciences, Boston, MA). Membranes were incubated for 2 h with a polyclonal anti-human CYP3A5 antibody specific for CYP3A5 (Affiniti, Exeter, UK) or a polyclonal CYP3A4 antibody reacting with CYP3A4, CYP3A5, and CYP3A7 (Daiichi, Tokyo, Japan). Blots were revealed after 4-chloronaphtol staining. After scanning, the band intensities were measured using the Densylab densitometric software (Microvision, Evry, France). Results were normalized using the intensity of the band from the same sample loaded onto each blot and expressed as optical density units per milligram of microsomal proteins.
Genomic DNA Preparation. Genomic DNA was purified from liver samples using the GenomicPrep Cells and Tissue DNA Isolation Kit (Amersham Biosciences UK, Ltd., Little Chalfont, Buckinghamshire, UK) according to the manufacturer's instructions, except for the rehydration step, in which sterile water was used instead of the provided solution so as not to interfere with subsequent amplification steps. The genomic DNA samples were stored at 4°C.
Primer Design and CYP3A5 Genotype Determination. Genbank accession number of sequences are the following: AC005020 .5 for CYP3A5; AF280107 .1 for CYP3A4, CYP3AP2, CYP3A7, CYP3AP1, and CYP3A43 promoter regions; M18907 .1, J04813 .1, D00408 .1, and AF319634 .1 for CYP3A4, CYP3A5, CYP3A7, and CYP3A43 mRNA, respectively; and X03205 .1 for 18S ribosomal RNA. Genomic and RNA sequences of CYP3A4, CYP3A5, CYP3A7, CYP3A43, CYP3AP2, and CYP3AP1 were aligned using the Multalin multiple sequence alignment software (Corpet, 1988). Oligonucleotides were selected using the Primer3 PCR primer design program (Rozen and Skaletsky, 2000) and were designed to ensure selectivity toward the CYP3A5 sequence. The sequence of oligonucleotides is shown in Table 1. Oligonucleotides for genotyping the CYP3A53 alleles (5020_22719, 5020_24161, 5020_22743, and 5020_23205) were designed by Kuehl et al. (2001), and CYP3A51/3 genotyping was performed as described previously.
Plasmid Construction. pMT2-3A5, a mammalian expression vector containing the CYP3A51 cDNA, was prepared previously (Santos et al., 2000). Oligonucleotides designed to hybridize to exons 2 and 8 were tailed with XhoI and KpnI without modification of the open reading frame. They were used to amplify the pMT2 sequence flanked with exons 1- to 2- and 8- to 13-spliced CYP3A5 exons. Genomic DNA from a 3/3 patient was amplified by high-fidelity PCR from exons 3 to 7 with oligonucleotides tailed with XhoI-modified exon 2 + 3 and KpnI-modified exon 7 + 8 sequences. The genomic insert and pMT2-flanked fragment were digested with XhoI and KpnI and ligated with T4 DNA ligase to obtain the p3A53C construct containing intronic and exonic sequences from exons 3 to 7 into 3A51 cDNA. The p3A53C minigene was entirely sequenced and presented the SNP configuration of the CYP3A53C allele (g.6986G and g.31611C). The structure of the minigene used herein is shown in Fig. 1. The GeneEditor in vitro Site-Directed Mutagenesis System (Promega, Madison, WI) was used to generate modified constructs. p3A53C-PTC was obtained after removal of the four premature termination codons present in exon 3B with oligonucleotide TGT GTC ACA CCC AAA CGA ACT AGA ACC AAA GGT TGC TGT GTG TCG TAC AAC AAG GGG TAT GGA TTA C. p3A53A was obtained after the mutation of p3A53C with oligonucleotide ATT CTA AGG ATT TCT ACT TTG and reproduces the CYP3A53A-like sequence (g.6986G and g.31611T).
Cell Culture. Human breast carcinoma MCF7 and human hepatocellular carcinoma HepG2 cells were grown at 37°C under 5% CO2, respectively, in Dulbecco's modified Eagle's medium and minimal essential medium supplemented with 10% fetal calf serum, 100 U/ml penicillin, 100 e蘥/ml streptomycin, and 1.5 e蘥/ml amphotericin B (Invitrogen, Cergy-Pontoise, France).
Cell Transfections. For the measurement of RNA decay, 5.105 MCF7 cells were seeded in six-well plates. After 24 h, cells were transfected using the calcium phosphate method with 5 e蘥 of minigene construct (p3A53C or p3A53C-PTC) and 2 e蘥 of pRL-SV40 (Promega). On day 3, transfected cells were incubated with 5 e蘥/ml actinomycin-D (Acros Organics, Geel, Belgium) for 2, 4, 6, and 8 h before RNA isolation with or without pretreatment with 20 e蘥/ml cycloheximide for 2 h (Acros Organics).
The effect of cycloheximide (20 e蘥/ml) on CYP3A5 mRNA level after mock, p3A53C, p3A53C-PTC, or p3A53A transfection was investigated in MCF7 cells after an incubation of 8 h using the same protocol, except that cells were not treated by actinomycin-D. The effect of cycloheximide (20 e蘥/ml) on the endogenous CYP3A5 mRNA level was also investigated in native HepG2 cells. Total RNA was extracted from frozen liver samples or cultured cells using the Promega SV Total RNA Isolation System.
siRNA Experiments. For siRNA experiments, MCF7 and HepG2 cells were seeded in 12-well plates and transfected on day 2 with 60 pmol of GL2 or UPF1 siRNAs using 3 e蘬 of Lipofectamine 2000 (Invitrogen) according to the manufacturer's recommendations. Annealed siRNA duplexes were purchased from MWG Biotech (Ebersberg, Germany); mRNA targets for gene-specific knockdown were AAC GUA CGC GGA AUA CUU CGA for the firefly luciferase gene (Elbashir et al., 2001) and AAG AUG CAG UUC CGC UCC AUU for the human UPF1 (RENT1) protein (Mendell et al., 2002). On day 3, cells were trypsinized and split at a low density into fresh 12-well plates, and transfection with 60 pmol of siRNA were repeated on day 6 using 3 e蘬 of oligofectamine (Invitrogen). Cells were trypsinized on day 7 and split in new 12-well plates before being transfected on day 8 with 100 ng of pGL2-pm (Promega), 50 ng of pRL, and 500 ng of either minigene construct with 3 e蘬 of Lipofectamine 2000. Cells were harvested on day 9, and RNA was extracted with TRIzol followed by digestion with RQ1 DNase (Promega) according to the manufacturer's recommendations. The luminescence emitted by the firefly luciferase synthesized from pGL2 and pRL was measured using the Dual-Glo Luciferase Assay System (Promega) on a Berthold Lumat LB 9507 luminometer.
The integrity of RNA was checked by electrophoresis in 1% agarose gels in 1x Tris-acetate/EDTA buffer. One microgram of total RNA was reverse-transcribed using 500 ng of random hexamers and 200 units of Moloney murine leukemia virus reverse transcriptase (Promega) in a 25-e蘬 final volume, according to the manufacturer's protocol.
PCR, Quantitative Real-Time PCR, and Data Normalization. Total RNA was reverse-transcribed and diluted 1:40 with PCR-grade water for the detection of CYP3A5 and GAPDH cDNA or with an additional 1:104 dilution for 18S rRNA cDNA detection. Quantitative real-time PCR experiments were performed with Lightcycler FastStart DNA Master Plus SYBR Green I (Roche Diagnostics GmbH, Mannheim, Germany), with oligonucleotides at a 300 nM final concentration in glass capillaries in a 10-e蘬 final volume. Total CYP3A5 mRNA was measured using oligonucleotides 2F and 3R, whereas splice variants (SV), which contain the pseudoexon 3B, were amplified using oligonucleotides MF and MR. CYP3A5 mRNAs correctly spliced at exon junctions 3 to 4 and 6 to 7, referred to as wild-type CYP3A5 mRNA (wt), were amplified using OKF and OKR. Polypyrimidine tract binding protein (PTB) mRNA were amplified by PCR with oligonucleotides 2132 and D3'2, described by Wollerton et al. (2004). These oligonucleotides allowed the amplification of a product spanning exons 10 to 12 and the detection of wild-type mRNA and splice variants lacking exon 11.
The hepatic CYP3A5 mRNA level was normalized with 18S rRNA, a suitable housekeeping gene for human liver samples (Koch et al., 2002). RNA decay experiments were normalized with 18S rRNA, which is transcribed with an actinomycin-D-insensitive polymerase. Q-PCR normalization was performed with the Renilla reniformis luciferase mRNA level in transiently transfected MCF7 and with endogenous GAPDH mRNA levels for experiments performed in HepG2 cells. In siRNA experiments, to allow a comparison between cell lines that present different levels of CYP3A5 expression, we expressed results as the ratio between 3B splice variants and wt CYP3A5 mRNA levels.
PCRs were performed with the PCR Core System I (Promega) with oligonucleotides 2F and 4R for the simultaneous amplification of wt and 3B variants. PCR products were analyzed by electrophoresis in 2.5% agarose gels.
Data Analysis. Statistical significance was assessed by Mann-Whitney U test or Student's t test using a spreadsheet program.
Results
Quantification of CYP3A5 mRNA in Human Livers. Twenty-seven samples were genotyped for the CYP3A53 allele, and the CYP3A5 mRNA levels was determined by quantitative real-time PCR (Fig. 2). It is noteworthy that in all samples bearing at least one CYP3A53 allele, the total RNA level is a composite figure of correctly spliced and splice variants in variable amounts. The CYP3A5 mRNA level is expressed in arbitrary units, with a value set to 1 for our single homozygous 1/1 liver. Livers from noninduced subjects were divided into subpopulations according to their genotype: CYP3A53/3 (n = 14), CYP3A51/3 (n = 7), and CYP3A51/1 (n = 1) carriers. The level of total CYP3A5 mRNA in the whole population varied over a 53-fold range. This variability fell to nine (range, 0.050-0.439) and five (range, 0.539-2.651) within CYP3A53/3 and 1/3 subpopulations, respectively. Ranges were not overlapping, and the difference between subgroups was statistically significant (Mann-Whitney U test, p < 0.001). When the proportion of correctly spliced CYP3A5 mRNA was examined, a significantly higher amount was observed (Mann-Whitney U test, p < 0.001) in CYP3A51/3 (range, 0.643-0.883) than in 3/3 carriers (range, 0.086-0.358), respectively, with a 1.5- and 4-fold intragroup variability. These results are consistent with the study of Lin et al. (2002), although the ratio of correctly spliced mRNA/splice-variant mRNA is different in 3/3 carriers.
It is interesting that the total CYP3A5 mRNA level is greater in 1 carriers than in 3/3 homozygotes. This information suggests either a difference in CYP3A5 mRNA stability or a reduction in the transcriptional activity of CYP3A53. So far, no SNP affecting the 5' regulatory region of the CYP3A5 gene has been reported to be in linkage disequilibrium with the CYP3A53 mutation, but this latter explanation cannot be excluded. Therefore, we wished to determine whether the mRNA stability could be modified in CYP3A5 splice variants that incorporate pseudoexon 3B.
CYP3A5 Wild-Type and Splice-Variant Decay. We developed two cellular models to explore the mechanism underlying CYP3A5 mRNA decay. First, the CYP3A5 genotype and mRNA expression level were investigated in the MCF7 human breast carcinoma and the HepG2 human hepatocellular carcinoma cell lines. These two cell lines have the CYP3A53/3 genotype. MCF7 cells present a very low basal CYP3A5 mRNA accumulation and constitute a suitable model for investigating the splicing and stability of CYP3A5 mRNA with an exogenous transfected CYP3A5 minigene construct. On the other hand, HepG2 cells express CYP3A5 mRNA at a higher basal level than MCF7 cells (>30-fold), allowing us to explore the splicing of the endogenous CYP3A5 gene (Fig. 3).
MCF7 cells were transfected with p3A53C to investigate the stability of wt and splice variants by quantitative real-time PCR after treatment with actinomycin-D, an inhibitor of RNA synthesis. Data indicated that splice variants containing the pseudoexon 3B decayed faster than the wild-type transcript (the half-life of wt mRNA is 1.7-fold longer than that of splice variants), suggesting that the difference in the total CYP3A5 mRNA level between 1 carriers and 3/3 carriers might be caused by a different stability of RNA species (Fig. 4).
Lee et al. (2003) suggested that the g.31611C>T mutation, which is in linkage disequilibrium with the g.6986A>G SNP, could be involved in the degradation of the CYP3A5 mRNAs. To evaluate the impact of this mutation on the decay of CYP3A5 mRNA in the MCF7 model, the level of CYP3A5 mRNA transcribed from p3A53A (generated by g.31611C>T site-directed mutagenesis of p3A53C) was compared with those produced by p3A53C 24 h after transfection with either construct (Fig. 5). Neither wild-type nor splice-variant CYP3A5 mRNA levels are modified by this SNP. Therefore, the total CYP3A5 mRNA level was unaffected, which would indicate that the g.31611C>T transition is unlikely to be implicated in the degradation of CYP3A5 mRNAs. In subsequent experiments, we used the p3A53C minigene to carry out transfection in MCF7 cells and investigate the mechanism of splice-variant degradation.
Protein Synthesis Inhibition Experiments. A possible mechanism for the degradation of mRNAs containing PTCs is NMD and involves ribosomes. Protein synthesis inhibitors have been widely used to demonstrate the role of NMD in the degradation of nonsense transcripts (Lei et al., 2001; Sureau et al., 2001; Lamba et al., 2003; Harries et al., 2004). Mock-transfected or MCF7 cells transfected with p3A53C were incubated or not with cycloheximide (Fig. 6). The level of wild-type and splice-variant CYP3A5 mRNA was more than 30-fold greater in minigene-transfected MCF7 than in native MCF7 cells (Student's t test, p < 0.001). Protein synthesis inhibition by cycloheximide resulted in an increase in the level of splice variant that contained the pseudoexon 3B (Student's t test, p < 0.001) (Fig. 6a). This effect was not caused by an induction of the endogenous CYP3A5 gene, as seen with mock-transfected MCF7 cells treated with cycloheximide. The treatment with cycloheximide has no significant effect on the amount of correctly spliced CYP3A5 mRNA, indicating that the effect might be specific to RNA molecules containing PTC (Fig. 6b). Furthermore, cycloheximide slows the decay of splice-variant RNA but has no effect on the decay of correctly spliced mRNA (data not shown). To confirm the effect of cycloheximide on the decay of splice variant, PCRs with different pairs of oligonucleotides were carried out with endogenously expressed CYP3A5 in HepG2 cells. As shown in Fig. 7b, the addition of cycloheximide has no effect on the level of wild-type RNA, whereas the level of splice variants is 2.5-fold higher than in cells treated with vehicle only (Fig. 7A). Thus, cycloheximide elicits the same effect on the endogenous CYP3A5 gene as in the transfected CYP3A53C minigene in MCF7, corresponding to a marked reduction of the decay of incorrectly spliced RNA, presumably through an inhibition of NMD.
Role of PTC in RNA Stability. NMD is a protection mechanism that promotes the degradation of mRNA containing PTC. The pseudoexon 3B is 132 base pairs long, and its insertion into the RNA molecule does not shift the ORF. Thus, mutation of PTC into amino acid coding sequences without modification of the g.6982A>G mutation should permit the synthesis of a splice-variant RNA not degraded by NMD. The site-directed mutagenesis of p3A53C was designed as p3A53C-PTC and was transfected into MCF7 cells. The amount of CYP3A5 splice variant was compared after transfection of p3A53C containing the sequence 3B and its four PTC or p3A53C-PTC containing the sequence 3B devoid of PTC. A marked increase (x3.8) in the level of splice variants containing pseudoexon 3B was detected with p3A53C-PTC (Fig. 8, p < 0.001) and was associated with a reduction of the decay of the transcript compared with splice variants produced from p3A53C (Fig. 4). This clearly indicates that the decay of splice variants is associated with the presence of PTCs and suggests the participation of NMD in the degradation of CYP3A53 splice variants.
Inhibition of hUPF1 by siRNA. NMD in mammalian cells depends on hUPF1 (or RENT1), an RNA-binding protein whose activity is modulated by ATP. Herein we have designed experiments to selectively knockdown hUPF1 with an siRNA described previously (Mendell et al., 2002). Crucial points are the use of adequate controls. To test the transfection efficiency with siRNAs, MCF7 cells were transfected with GL2 siRNA directed against the R. reniformis luciferase mRNA: this resulted in an 85 ± 5% (mean ± S.E.M.) inhibition of the luciferase activity, whereas transfection with UPF1 siRNA had no effect on the luciferase activity. To test the transfection with siRNA for hUPF1 in MCF7 or HepG2 cells, cellular RNA was extracted and used as a template in reverse-transcriptase PCR for the amplification of the PTB mRNA. A 30 ± 3% (mean ± S.E.M.) increase in PTB splice variant lacking exon 11, a known substrate of NMD (Wollerton et al., 2004), was noticed, indicating that NMD activity was reduced. Mock-transfected cells and cells transfected with GL2 siRNA showed no variation in the level of PTB splice variant. Thus, hUPF1 siRNA acts specifically on the NMD activity.
The level of endogenous CYP3A5 splice variant mRNA is twice as high in HepG2 cells transfected with UPF1 siRNA as in mock-transfected or HepG2 cells transfected with GL2 control siRNA (Fig. 9a). Similar results were obtained in MCF7 cells transfected with exogenous CYP3A53C plasmid in the absence or presence of hUPF1 siRNA (Fig. 9b). These results confirm the involvement of NMD in the degradation of CYP3A53 splice variant containing the pseudoexon 3B and its four PTCs and explain the rapid decay of CYP3A53 RNA.
Inducibility of CYP3A5. It has been shown recently that the addition of CYP3A inducers to human hepatocytes could substantially increase the CYP3A5 mRNA level (Burk et al., 2004). We wished to determine whether the administration of barbiturates or steroids to patients at pharmacological doses could modify the splice variant/wild-type RNA ratio, shifting the 3/3 genotype to a 1/3-like phenotype. None of our adult liver samples were medicated and so did not allow for this type of comparison. However, in our pediatric liver bank, five samples were obtained from neonates or newborns treated with barbiturates or steroids.
Compared with untreated samples in the same age group, samples treated with barbiturates displayed a higher CYP3A content associated with an enhanced testosterone-6 hydroxylation (Table 2). In a child treated for 1 year with prednisone, no increase was noticed. When tested with an antibody specific for CYP3A5, only N77 was positive, but his CYP3A5 mRNA content was comparable with that of age-matched newborns receiving no treatment. The four other samples had very low levels of CYP3A5 corresponding to a 3/3 phenotype. The level of CYP3A5 mRNA and the proportion of wild-type CYP3A5 mRNA were investigated in these samples. In essence, the total CYP3A5 mRNA was comparable in pediatric and adult populations (Fig. 2). When classified according to their genotype, 3/3 neonates exhibited a CYP3A5 RNA level close to that of 3/3 adults; 1/3 neonates were close to 1/3 adults, and the proportion of correctly spliced RNA was comparable in pediatric and adult populations according to their genotype. Thus, no age-dependent effect could be demonstrated with these samples. When "induced" samples were examined, the values obtained with the four 3/3 pediatric samples did not differ from those obtained with untreated 3/3 children or adults (Fig. 2). Their total RNA content and their proportion of correctly spliced RNA were similar to those of untreated age-matched subjects. N77, with a 1/3 genotype and induced by phenobarbital, displayed an unexpectedly low CYP3A5 mRNA content and a depressed proportion of wild-type correctly spliced mRNA in the range of 3/3 carriers. Therefore, our results do not support a potent CYP3A5 induction in five pediatric patients treated with barbiturates or steroids, although their total CYP3A content was markedly increased. Furthermore, treatment with a barbiturate did not modify the proportion of splice-variant mRNA nor the apparent decay of differently spliced RNA.
Discussion
The CYP3A5 protein is found at an appreciable level in only 10 to 20% of the white population. In the last decade, several hypotheses have been formulated to explain the polymorphism in CYP3A5 expression: 1) an instability of the newly synthesized protein; 2) a reduced transcriptional level caused by mutations in the promoter region; and 3) an altered level of mature RNA coding for the full-length protein.
First, Jounaidi et al. (1996) proposed that an SNP (CYP3A52 allele) could make the CYP3A5 protein unstable. However, the very low frequency of this allele (0.019) (Hustert et al., 2001) could not explain the low level of protein in 80% of the white population. Moreover, recombinant T398N mutant protein is detected by Western blotting in microsomes from cells transfected with the CYP3A52 cDNA at the same level as after transfection with wild-type CYP3A5 cDNA and therefore does not seem to be unstable (F. Busi, G. Aubert, and T. Cresteil, unpublished data). Later, Paulussen et al. (2000) were concerned with an alteration of the transcriptional level of CYP3A5 and reported a good correlation between the CYP3A5 expression and the presence of two mutations in what was believed to be the CYP3A5 promoter region and was, in fact, located upstream of the CYP3AP1 pseudogene. To date, no SNP correlated with CYP3A5 expression has been described in the promoter region of CYP3A5, and no evidence has been provided showing a reduced transcription of CYP3A5 in a fraction of the population.
Kuehl et al. (2001) produced convincing evidence that the low level of CYP3A5 protein was associated with the presence of two CYP3A53 alleles. The g.6986A>G mutation produced an alteration of the CYP3A5 splicing with the insertion of pseudoexon 3B. This mutation resulted in a lower RNA content and was tentatively explained by a rapid degradation of splice variant by NMD. NMD is a protective process for the cell promoting degradation of mRNAs containing premature termination codons located more than 50 nucleotides before the last exon-exon junction (Holbrook et al., 2004). With four PTCs, splice variants of CYP3A53 containing the pseudoexon 3B are possible substrates for NMD. However this hypothesis has not been verified so far. The last hypothesis proposed by Lee et al. (2003) implicated the g.31611C>T mutation. This mutation is in linkage disequilibrium with the g.6986A>G SNP. The g.31611C>T SNP creates an AUUUC motif in the 3' untranslated region (3'-UTR) of CYP3A5 mRNAs produced from CYP3A53 alleles. This sequence resembles the AUUUA 3'-UTR destabilizing motif observed in some highly regulated mRNAs (Wilusz et al., 2001).
The goal of the present study was to gain insight into the molecular mechanism controlling CYP3A5 expression. For that purpose, two models were designed to test results. A "minigene" was built containing introns 3 to 6 to generate splice variants. This construction was transfected into MCF7 cells, a cell line having an extremely low basal expression of the endogenous CYP3A5 gene. Data obtained after transfection experiments were compared with those of the endogenous CYP3A5 expression in HepG2, a cell line with an active CYP3A53 gene transcription, generating splice variants. The initial observation was related to the decay of CYP3A5 mRNA in cell lines: splice variants which incorporate the pseudoexon 3B are less stable than wild-type mRNA: this clearly indicates that in subjects bearing the 3/3 genotype, the lowest CYP3A5 mRNA level could result from a rapid degradation of splice variants produced by CYP3A53.
In a first attempt to tentatively explain the degradation of splice-variant RNA, the role of the g.31611C>T mutation was investigated. Reversal of this mutation to the wt genotype had no effect on the decay of splice variants containing the g.6986A>G mutation. This clearly indicates that the mutation in the 3'-UTR does not cause the faster degradation of the splice-variant RNA. On the other hand, the presence of multiple premature stop codons could suggest the implication of NMD in the RNA degradation. To test this hypothesis, cells were treated with cycloheximide, a molecule known to prevent protein synthesis and, indirectly, the NMD-directed RNA degradation. The addition of cycloheximide resulted in a marked accumulation of splice variants through an inhibition of its fast degradation but has no effect on the degradation of the wt RNA. Therefore, a functional translation machinery is required for the degradation of splice-variant RNA. The removal of PTC from splice variants significantly reduced the degradation of splice variants, indicating the role of PTC in the degradation process through an activation of NMD. Finally, the silencing of hUPF1 by siRNA down-regulated the NMD pathway and was accompanied by a reduction of splice-variant degradation.
Altogether, these experiments conducted with both the exogenous minigene transfected in MCF7 cells and the endogenous CYP3A5 gene expressed in HepG2 cells strongly suggested the participation of NMD in the degradation of splice variants, whereas the wt CYP3A5 RNA is not affected.
In agreement with our in vitro cellular models, published results obtained in vivo from liver samples support the existence of an allele-specific degradation of CYP3A5 mRNA. In heterozygous g.31611C/T samples, only cDNA derived from the sequence containing the cytosine was detected and is called pseudohomozygocity by Wojnowski and Brockmoller (2004). The authors attribute this over representation to the tight linkage of g.31611C with g.6986A. These observations could be interpreted as a lower accumulation of CYP3A53 RNA than CYP3A51 RNA consequent to the rapid degradation of CYP3A53 mRNA containing PTC by NMD, in agreement with our observations.
Burk et al. (2004) have described a pregnane X receptor-dependent induction of CYP3A5 that could phenocopy the effect of the high-expression allele 1 in subjects bearing the 3/3 genotype. It could be speculated that the overall mRNA synthesis is stimulated, keeping constant the proportion of correctly and incorrectly spliced molecules synthesized. Because the splice variant is more rapidly degraded than the correctly spliced molecule, this should result in a higher content of wt RNA molecule and consequently in a higher protein synthesis. However, in our hands, no such observations have been validated: in 3/3 genotyped samples, the proportion of splice-variant/wild-type mRNA remains constant regardless of the induction by phenobarbital or prednisone, and no accumulation of CYP3A5 protein was noticed in contrast to the increased CYP3A (presumably CYP3A4) and related activity of testosterone-6-hydroxylase. These results have been obtained in pediatric patients treated with therapeutic doses of phenobarbital. In our samples, the expression of CYP3A5 in the white population was conserved in neonatal and adult groups with approximately 15% positive CYP3A5 samples, and the ontogenic pattern of CYP3A5 follows that described for CYP3A4 in the same liver bank (T. Cresteil and F. Busi, unpublished data). Stevens et al. (2003) also detected immunoreactive CYP3A5 protein in postnatal samples. Although the number of pretreated human samples was very limited, our data do not support the existence of a significant induction of hepatic CYP3A5 at this stage of development. At present, we cannot provide convincing evidence to explain the failure of phenobarbital to induce CYP3A5.
In conclusion, we confirm an observation published previously reporting a lower level of CYP3A5 mRNA in CYP3A53/3 carriers compared with CYP3A51 carriers. The splice variants generated by the CYP3A53 allele contained multiple premature stop codons and are degraded by nonsense-mediated mRNA decay faster than the wild-type mRNA, which explains the difference in mRNA levels observed between the two genotype subgroups. In pediatric samples obtained from barbiturate-treated children, the level of total CYP3A5 RNA and the proportion of splice-variant mRNA were not modified and correlated with the low level of CYP3A5 protein.
Acknowledgements
We are grateful to Michelle Sonnier for her helpful work in molecular biology and cell culture, to Dr. Alain Sureau and the group of Dr. Jolle Marie for their expert advices on pre-mRNA splicing, and to Dr. Roger Karess for improving the article.
References
Burk O, Koch I, Raucy J, Hustert E, Eichelbaum M, Brockmoller J, Zanger UM, and Wojnowski L (2004) The induction of cytochrome P450 3A5 (CYP3A5) in the human liver and intestine is mediated by the xenobiotic sensors pregnane X receptor (PXR) and constitutively activated receptor (CAR). J Biol Chem 279: 38379-38385.
Corpet F (1988) Multiple sequence alignment with hierarchical clustering. Nucleic Acids Res 16: 10881-10890.
Cresteil T, Flinois JP, Pfister A, and Leroux JP (1979) Effect of microsomal preparations and induction on cytochrome P-450-dependent monooxygenases in fetal and neonatal rat liver. Biochem Pharmacol 28: 2057-2063.
Cresteil T, Monsarrat B, Alvinerie P, Treluyer JM, Vieira I, and Wright M (1994) Taxol metabolism by human liver microsomes: identification of cytochrome P450 isozymes involved in its biotransformation. Cancer Res 54: 386-392.
Domanski TL, Finta C, Halpert JR, and Zaphiropoulos PG (2001) cDNA cloning and initial characterization of CYP3A43, a novel human cytochrome P450. Mol Pharmacol 59: 386-392.
Eichelbaum M and Burk O (2001) CYP3A Genetics in drug metabolism. Nat Med 7: 285-287.
Elbashir SM, Harborth J, Lendeckel W, Yalcin A, Weber K, and Tuschl T (2001) Duplexes of 21-nucleotide RNAs mediate RNA interference in cultured mammalian cells. Nature (Lond) 411: 494-498.
Fukuda T, Onishi S, Fukuen S, Ikenaga Y, Ohno M, Hoshino K, Matsumoto K, Maihara A, Momiyama K, Ito T, et al. (2004) CYP3A5 genotype did not impact on nifedipine disposition in healthy volunteers. Pharmacogenomics J 4: 34-39.
Gellner K, Eiselt R, Hustert E, Arnold H, Koch I, Haberl M, Deglmann CJ, Burk O, Buntefuss D, Escher S, et al. (2001) Genomic organization of the human CYP3A locus: identification of a new, inducible CYP3A gene. Pharmacogenetics 11: 111-121.
Harries LW, Hattersley AT, and Ellard S (2004) Messenger RNA transcripts of the hepatocyte nuclear factor-1alpha gene containing premature termination codons are subject to nonsense-mediated decay. Diabetes 53: 500-504.
Hesselink DA, van Schaik RH, van der Heiden I, van der Werf M, Gregoor PJ, Lindemans J, Weimar W, and van Gelder T (2003) Genetic polymorphisms of the CYP3A4, CYP3A5 and MDR-1 genes and pharmacokinetics of the calcineurin inhibitors cyclosporine and tacrolimus. Clin Pharmacol Ther 74: 245-254.
Holbrook JA, Neu-Yilik G, Hentze MW, and Kulozik AE (2004) Nonsense-mediated decay approaches the clinic. Nat Genet 36: 801-808.
Huang W, Lin YS, McConn DJ, Calamia JC, Totah RA, Isoherranen N, Glodowski M, and Thummel KE (2004) Evidence of significant contribution from CYP3A5 to hepatic drug metabolism. Drug Metab Dispos 32: 1434-1445.
Hustert E, Haberl M, Burk O, Wolbold R, He YQ, Klein K, Nuessler AC, Neuhaus P, Klattig J, Eiselt R, et al. (2001) The genetic determinants of the CYP3A5 polymorphism. Pharmacogenetics 11: 773-779.
Jounaidi Y, Hyrailles V, Gervot L, and Maurel P (1996) Detection of CYP3A5 allelic variant: a candidate for the polymorphic expression of the protein Biochem Biophys Res Commun 221: 466-470.
Katz DA, Grimm DR, Cassar SC, Gentile MC, Ye X, Rieser MJ, Gordon EF, Polzin JE, Gustavson LE, Driscoll RM, et al. (2004) CYP3A5 genotype has a dose-dependent effect on ABT-773 plasma levels. Clin Pharmacol Ther 75: 516-528.
Koch I, Weil R, Wolbold R, Brockmoller J, Hustert E, Burk O, Nuessler A, Neuhaus P, Eichelbaum M, Zanger U, et al. (2002) Interindividual variability and tissue-specificity in the expression of cytochrome P450 3A mRNA. Drug Metab Dispos 30: 1108-1114.
Kuehl P, Zhang J, Lin Y, Lamba J, Assem M, Schuetz J, Watkins PB, Daly A, Wrighton SA, Hall SD, et al. (2001) Sequence diversity in CYP3A promoters and characterization of the genetic basis of polymorphic CYP3A5 expression. Nat Genet 27: 383-391.
Lacroix D, Sonnier M, Moncion A, Cheron G, and Cresteil T (1997) Expression of CYP3A in the human liver—evidence that the shift between CYP3A7 and CYP3A4 occurs immediately after birth. Eur J Biochem 247: 625-634.
Lamba JK, Adachi M, Sun D, Tammur J, Schuetz EG, Allikmets R, and Schuetz JD (2003) Nonsense mediated decay downregulates conserved alternatively spliced ABCC4 transcripts bearing nonsense codons. Hum Mol Genet 12: 99-109.
Lee SJ, Usmani KA, Chanas B, Ghanayem B, Xi T, Hodgson E, Mohrenweiser HW, and Goldstein JA (2003) Genetic findings and functional studies of human CYP3A5 single nucleotide polymorphisms in different ethnic groups. Pharmacogenetics 13: 461-472.
Lei XD, Chapman B, and Hankinson O (2001) Loss of Cyp1a1 messenger RNA expression due to nonsense-mediated decay. Mol Pharmacol 60: 388-393.
Lin YS, Dowling AL, Quigley SD, Farin FM, Zhang J, Lamba J, Schuetz EG, and Thummel KE (2002) Coregulation of CYP3A4 and CYP3A5 and contribution to hepatic and intestinal midazolam metabolism. Mol Pharmacol 62: 162-172.
Macphee IA, Fredericks S, Tai T, Syrris P, Carter ND, Johnston A, Goldberg L, and Holt DW (2002) Tacrolimus pharmacogenetics: polymorphisms associated with expression of cytochrome P4503A5 and P-glycoprotein correlate with dose requirement. Transplantation 74: 1486-1489.
Mendell JT, ap Rhys CM, and Dietz HC (2002) Separable roles for Rent1/hUpf1 in altered splicing and decay of nonsense transcripts. Science (Wash DC) 298: 419-422.
Oscarson M and Ingelman-Sundberg M (2002) CYPalleles: a web page for nomenclature of human cytochrome P450 alleles. Drug Metab Pharmacokinet 17: 491-495.
Paulussen A, Lavrijsen K, Bohets H, Hendrickx J, Verhasselt P, Luyten W, Konings F, and Armstrong M (2000) Two linked mutations in transcriptional regulatory elements of the CYP3A5 gene constitute the major genetic determinant of polymorphic activity in humans. Pharmacogenetics 10: 415-424.
Rozen S and Skaletsky H (2000) Primer3 on the WWW for general users and for biologist programmers. Methods Mol Biol 132: 365-386.
Santos A, Zanetta S, Cresteil T, Deroussent A, Pein F, Raymond E, Vernillet L, Risse ML, Boige V, Gouyette A, et al. (2000) Metabolism of irinotecan (CPT-11) by CYP3A4 and CYP3A5 in humans. Clin Cancer Res 6: 2012-2020.
Shih PS and Huang JD (2002) Pharmacokinetics of midazolam and 1'-hydroxymidazolam in Chinese with different CYP3A5 genotypes. Drug Metab Dispos 30: 1491-1496.
Stevens JC, Hines RN, Gu C, Koukouritaki SB, Manro JR, Tandler PJ, and Zaya MJ (2003) Developmental expression of the major human hepatic CYP3A enzymes. J Pharmacol Exp Ther 307: 573-582.
Sureau A, Gattoni R, Dooghe Y, Stevenin J, and Soret J (2001) SC35 autoregulates its expression by promoting splicing events that destabilize its mRNAs. EMBO (Eur Mol Biol Organ) J 20: 1785-1796.
Westlind-Johnsson A, Malmebo S, Johansson A, Otter C, Andersson TB, Johansson I, Edwards RJ, Boobis AR, and Ingelman-Sundberg M (2003) Comparative analysis of CYP3A expression in human liver suggests only a minor role for CYP3A5 in drug metabolism. Drug Metab Dispos 31: 755-761.
Wilusz CJ, Wormington M, and Peltz SW (2001) The cap-to-tail guide to mRNA turnover. Nat Rev Mol Cell Biol 2: 237-246.
Wojnowski L and Brockmoller J (2004) Single nucleotide polymorphism characterization by MRNA expression imbalance assessment. Pharmacogenetics 14: 267-269.
Wollerton MC, Gooding C, Wagner EJ, Garcia-Blanco MA, and Smith CW (2004) Autoregulation of polypyrimidine tract binding protein by alternative splicing leading to nonsense-mediated decay. Mol Cell 13: 91-100.
Wrighton SA, Ring BJ, Watkins PB, and VandenBranden M (1989) Identification of a polymorphically expressed member of the human cytochrome P-450III family. Mol Pharmacol 36: 97-105.
Zheng H, Webber S, Zeevi A, Schuetz E, Zhang J, Bowman P, Boyle G, Law Y, Miller S, Lamba J, et al. (2003) Tacrolimus dosing in pediatric heart transplant patients is related to CYP3A5 and MDR1 gene polymorphisms. Am J Transplant 3: 477-483.