Post-transcriptional Regulation of Human Pregnane X Receptor by Micro-RNA Affects the Expression of Cytochrome P450 3A4*

Pregnane X receptor (PXR) is a major transcription factor regulating the inducible expression of a variety of transporters and drug-metabolizing enzymes, including CYP3A4 (cytochrome P450 3A4). We first found that the PXR mRNA level was not correlated with the PXR protein level in a panel of 25 human livers, indicating the involvement of post-transcriptional regulation. Notably, a potential miR-148a recognition element was identified in the 3′-untranslated region of human PXR mRNA. We investigated whether PXR might be regulated by miR-148a. A reporter assay revealed that miR-148a could recognize the miR-148a recognition element of PXR mRNA. The PXR protein level was decreased by the overexpression of miR-148a, whereas it was increased by inhibition of miR-148a. The miR-148a-dependent decrease of PXR protein attenuated the induction CYP3A4 mRNA. Furthermore, the translational efficiency of PXR (PXR protein/PXR mRNA ratio) was inversely correlated with the expression levels of miR-148a in a panel of 25 human livers, supporting the miR-148a-dependent regulation of PXR in human livers. Eventually, the PXR protein level was significantly correlated with the CYP3A4 mRNA and protein levels. In conclusion, we found that miR-148a post-transcriptionally regulated human PXR, resulting in the modulation of the inducible and/or constitutive levels of CYP3A4 in human liver. This study will provide new insight into the unsolved mechanism of the large interindividual variability of CYP3A4 expression.

A key function of the liver is the metabolism and elimination of xenobiotics or endobiotics. The expression of genes involved in these processes is largely regulated by transcription factors belonging to the nuclear receptor family. Pregnane X receptor (PXR 2 ; alternate names SXR, PAR, and NR1I2), a member of the nuclear receptor family, is a crucial regulator of drug metabolism and elimination. It is predominantly expressed in liver and small intestine. PXR is activated by a broad spectrum of xenobiotics, including antibiotics, antimycotics, and herbal components (1); dimerizes with retinoid X receptor ␣ (RXR␣); and binds to response elements of target genes, including cytochrome P450s, UDP-glucuronosyltransferases, glutathione S-transferases, sulfotransferases, and various transporters, such as MDR1 (multidrug resistance 1) and MRP-2 (multidrug resistance-associated protein 2) to induce them (2). Thus, PXR is recognized as a xenosensor for the detoxification of foreign compounds. However, it also plays a role as a physiological sensor of bile acids to protect the body from toxicity by regulating the expression of target genes that decreases the synthesis and increases the elimination of bile acids (3,4).
One of the best known genes regulated by PXR is CYP3A4, the most abundant P450 in human liver that catalyzes the metabolism of over 50% of current prescription drugs (5)(6)(7). A large interindividual difference (ϳ50-fold) has been reported for the CYP3A4 level in the general population (8), which cannot be explained by genetic polymorphisms (9,10). The CYP3A4 expression is largely regulated at the transcriptional level by transcriptional factors, such as CCAAT/enhancerbinding proteins, C/EBP␣ and C/EBP␤, and hepatocyte nuclear factors, HNF4␣ and HNF3␥, as well as constitutive androstane receptor (CAR) and PXR (11). However, the cause of the large interindividual variability in CYP3A4 level is poorly understood and is an urgent issue to be solved. The regulation by PXR may, in part, be responsible for such variability, since PXR is activated by endogenous compounds, such as steroid hormones and bile acids (1,12).
PXR regulates many targets controlling pharmacokinetics, but its own regulation is not fully understood, with reports showing only that human PXR is induced by dexamethasone through glucocorticoid receptor (13) or by clofibrate through peroxisome proliferator-activated receptor ␣ (14). Employing an on-line search using the miRBase Target data base (15) (available on the World Wide Web), we found some potential recognition sites for micro-RNAs (miRNAs) in the 3Ј-untranslated region (UTR) of the human PXR.
miRNAs are a recently discovered family of short noncoding RNA whose final product is an ϳ22-nucleotide functional RNA molecule (16). They play important roles in the regulation of target genes by binding to complementary regions of transcripts to repress their translation or regulate degradation. At present, more than 400 miRNAs have been identified in humans, and miRNAs are predicted to control about 30% of the genes within the human genome (17,18). The roles of miRNAs have received attention especially in the cancer field but hardly yet in the field of pharmacokinetics. In the present study, we investigated whether human PXR might be post-transcriptionally regulated by miRNA and its impact on CYP3A4 expression.
Human Livers and Cell Culture Conditions-Human liver samples from 25 donors were obtained from the Human and Animal Bridging Research Organization (Chiba, Japan). The human hepatocellular carcinoma cell lines HepG2 and HuH7 were obtained from Riken Gene Bank (Tsukuba, Japan), and HLE was from the Japanese Collection of Research Bioresources (Tokyo, Japan). The human colon carcinoma cell lines LS180 and Caco-2, the human embryonic kidney cell line HEK293, and the human breast adenocarcinoma cell line MCF-7 were obtained from American Type Culture Collection (Manassas, VA). HepG2, HuH7, and HLE cells were cultured in Dulbecco's modified Eagle's medium (Nissui Pharmaceutical, Tokyo, Japan) supplemented with 10% fetal bovine serum (Invitrogen). LS180, Caco-2, and MCF-7 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 0.1 mM nonessential amino acid (Invitrogen) and 10% fetal bovine serum. Differentiated Caco-2 (Caco-2/D) cells were obtained by culture for 3 weeks postconfluence. HEK293 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 4.5 g/liter glucose, 10 mM HEPES, and 10% fetal bovine serum. These cells were maintained at 37°C under an atmosphere of 5% CO 2 , 95% air.
Real Time RT-PCR for PXR and CYP3A4-Total RNA was isolated from 25 human liver samples using ISOGEN (Nippon Gene, Tokyo, Japan) according to the manufacturer's protocol. The cDNAs were synthesized from total RNA using ReverTra Ace (Toyobo, Osaka, Japan). The forward and reverse primers for CYP3A4 were 5Ј-CCAAGCTATGCTCTTCACCG-3Ј and 5Ј-TCAGGCTCCACTTACGGTGC-3Ј, respectively. The forward and reverse primers for human PXR were 5Ј-TGCGAG-ATCACCCGGAAGAC-3Ј and 5Ј-ATGGGAGAAGGTAGT-GTCAAAGG-3Ј, respectively. The real time PCR was performed using the Smart Cycler (Cepheid, Sunnyvale, CA) with Smart Cycler software (version 1.2b) as follows. After an initial denaturation at 95°C for 30 s, the amplification was performed by denaturation at 95°C for 6 s and annealing and extension at 68°C for 20 s for 40 cycles. The mRNA levels were normalized with GAPDH mRNA determined by real time RT-PCR, as described previously (19).
SDS-PAGE and Western Blot Analyses of PXR and CYP3A4-Whole cell lysates were prepared from 25 human liver samples by homogenization with lysis buffer (50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40) containing protease inhibitors (0.5 mM (p-amidinophenyl)methanesulfonyl fluoride, 2 g/ml aprotinin, 2 g/ml leupeptin). The protein concentrations were determined using Bradford protein assay reagent (Bio-Rad) with ␥-globulin as a standard. The whole cell lysates (10 -50 g) were separated with 7.5% SDS-polyacrylamide gel electrophoresis and transferred to Immobilon-P transfer membrane (Millipore, Bedford, MA). The membranes were probed with goat anti-human PXR, rabbit anti-human CYP3A4, goat anti-human HNF4␣, rabbit anti-human CAR, or rabbit antihuman GAPDH antibodies and the corresponding fluorescent dye-conjugated second antibody, and the band densities were quantified with an Odyssey infrared imaging system (LI-COR Biosciences). Nuclear extracts (10 g) from the HepG2 and LS-180 cells were also used to determine the PXR protein level.
Real Time RT-PCR for Mature miR-148a-For the quantification of mature miR-148a, polyadenylation and reverse transcription were performed using the NCode miRNA first strand cDNA synthesis kit (Invitrogen) according to the manufacturer's protocol. The forward primer for miR-148a was 5Ј-TCAG-TGCACTACAGAACTTTGT-3Ј, and the reverse primer was the supplemented universal qPCR primer. The PCR analyses were performed as follows. After an initial denaturation at 95°C for 30 s, the amplification was performed by denaturation at 95°C for 10 s and annealing and extension at 64°C for 10 s for 40 cycles. The mature miR-148a level was normalized with U6 snRNA determined by real time RT-PCR, as described previously (20).
Construction of Reporter Plasmids-To construct luciferase reporter plasmids, various target fragments were inserted into the XbaI site downstream of the luciferase gene in the pGL3promoter vector. The sequence from 3362 to 3383 in the human PXR mRNA (5Ј-ACAGACTCTTACGTGGAGAGTG-CACTGA-3Ј) was termed the miR-148a recognition element (PXRMRE148). The fragment containing three copies of the PXRMRE148 (5Ј-CTAGAAGCCACAGACTCTTACGTGGAG-AGTGCACTGACCTGTAGAAGCCACAGACTCTTACGTG-GAGAGTGCACTGACCTGTAGAAGCCACAGACTCTTAC-GTGGAGAGTGCACTGACCTGTAT-3Ј; PXRMRE148 is underlined) was cloned into the pGL3-promoter vector (pGL3p/3ϫPXRMRE). The complementary sequence of three copies of the PXRMRE148 was also cloned into the pGL3-promoter plasmid (pGL3p/3ϫPXRMRE-Rev). A fragment containing the perfect matching sequence with the mature miR-148a, 5Ј-CTAGACAAAGTTCTGTAGTGCACTGAT-3Ј (the matching sequence of miR-148a is underlined) was cloned into the pGL3-promoter vector (pGL3p/c-148a). The nucleotide sequences of the constructed plasmids were confirmed by DNA sequencing analyses.
Luciferase Assay-Various luciferase reporter plasmids (pGL3p) were transiently transfected with pGL4.74-TK plasmid into HEK293 or HepG2 cells. Briefly, the day before transfection, the cells were seeded into 24-well plates. After 24 h, 70 ng of pGL3p plasmid, 30 ng of pGL4.74-TK plasmid, and 4 pmol of the precursors for miR-148a or control were transfected into HEK293 cells using Lipofectamine 2000. For HepG2 cells, 80 ng of pGL3p plasmid, 20 ng of pGL4.74-TK plasmid, and 10 pmol of the AsOs for miR-148a or control were transfected using Tfx-20 reagent. After incubation for 48 h, the cells were resuspended in passive lysis buffer, and then the luciferase activity was measured with a luminometer (Wallac, Turku, Finland), using the dual luciferase reporter assay system.

Transfection of Precursor or Antisense for miR-148a into HepG2 and LS-180 Cells and Isolation of Nuclear Extract and
Total RNA-To investigate the effect of miR-148a on the expression of PXR protein, 50 nM precursor or 50 nM AsO for miR-148a or control was transfected into HepG2 cells using Lipofectamine RNAiMAX. After 72 h, total RNA was isolated using ISOGEN, and the mature miR-148a levels were determined by Northern blotting as described above. Nuclear extract was isolated using the NE-PER nuclear and cytoplasmic extraction reagents (Pierce) according to the manufacturer's protocol. LS180 cells were transfected with 50 nM precursor for miR-148a or control using Lipofectamine RNAiMAX. After 72 h, the cells were treated with 50 M rifampicin or 0.1% (v/v) Me 2 SO for 24 h. Then total RNA and nuclear extract were isolated.
Evaluation of the Expression Level of PXR in HepG2 Using Reporter Construct Containing PXR-responsive Element-The reporter construct pCYP3A4-362-7.7K contains the promoter region (Ϫ362 to ϩ11), including the ER6 (everted repeat separated by six nucleotides) motif and the distal enhancer region (Ϫ7836 to Ϫ7200), including the DR3 (direct repeat separated by three nucleotides) motif of the CYP3A4 gene, to which PXR binds (21). The day before transfection, HepG2 cells were seeded into 24-well plates. After 24 h, 180 ng of pCYP3A4-362-7.7K, 20 ng of pGL4.74-TK plasmid, and various doses of the precursors and AsOs for miR-148a or control were transfected using Tfx-20 reagent. After incubation for 48 h, the cells were treated with 10 M rifampicin or 0.1% Me 2 SO for 24 h, and then the luciferase activity was measured.
Statistical Analyses-Statistical significance was determined by analysis of variance followed by Dunnett's multiple comparisons test or Tukey's method test. Comparison of two groups was made with an unpaired, two-tailed Student's t test. Corre-lation analyses were performed by Spearman's rank method. A value of p Ͻ 0.05 was considered statistically significant.

PXR Protein Level Is Not Associated with PXR mRNA Level in
Human Livers-We first examined the PXR mRNA level in a panel of 25 human livers by real time RT-PCR assay and investigated the relationship with the PXR protein level. As shown in Fig. 1, no statistically significant correlation was observed between the PXR mRNA and protein levels (Rs ϭ 0.10), indicating the involvement of post-transcriptional regulation of human PXR. To uncover the molecular mechanism of the posttranscriptional regulation, we sought to examine the involvement of miRNA-mediated regulation. Employing an on-line search using the miRBase Target data base (15) (available on the World Wide Web), potential recognition elements for 16 kinds of miRNA, such as miR-148a, miR-192, and miR-560, were found in the 3Ј-UTR in human PXR. Among them, we focused on miR-148a because it is selectively and abundantly expressed in liver (22) and has high complementarity in the 5Ј-end at miRNA-mRNA duplexes, including the seed sequence. The potential miR-148a target site is ϳ200 bases downstream of the stop codon of the human PXR mRNA. The alignment of hsa-miR-148a with the 3Ј-UTR of human PXR mRNA ( Fig. 2A) was drawn using RNAhybrid (23) (available on the World Wide Web). We investigated whether this region, termed the miR-148a recognition element (PXRMRE148), might be involved in the regulation of PXR by miR-148a.
Expression Levels of miR-148a in Human Cancer Cell Lines-Real time RT-PCR analysis using the NCode miRNA first strand cDNA synthesis kit was performed to determine the expression levels of mature miR-148a in eight kinds of human cancer cell lines (Fig. 2B). The mature miR-148a was detected in all cell lines tested in this study, with large variability among cell lines (37-fold). The U6 snRNA levels, which were used for nor- malization, varied Ͻ3-fold in the experiment, much less than the miR-148a levels. The mature miR-148a was higher in HepG2 and differentiated Caco-2 cells than the other cell lines. It was also interesting that, in Caco-2 cells, the expression level of mature miR-148a was increased with differentiation. Thus, the expression levels of mature miR-148a were highly variable among the human cancer cell lines.
Repressive Regulation of PXR by miR-148a in Human Cell Lines-To investigate whether PXRMRE148 is functional in the regulation by miR-148a, luciferase assays were performed with HEK293 cells (Fig. 2C). We first confirmed that the luciferase activity of the pGL3p/c-148a plasmid, in which the miR-148a complementary sequence was inserted downstream of the luciferase gene, was significantly (p Ͻ 0.01) decreased by the co-transfection with the precursor for miR-148a. The luciferase activity of pGL3p/ 3ϫ PXRMRE plasmid, in which three copies of the potential miR-148a recognition site were inserted downstream of the luciferase gene, was also significantly (p Ͻ 0.01) decreased by co-transfection with the precursor for miR-148a (33% of control), whereas that of pGL3p/ 3ϫ PXRMRE-Rev plasmid with the inverted recognition site was not affected. In HepG2 cells, which showed the highest expression of miR-148a (Fig. 2B), the luciferase activities of pGL3p/c-148a and pGL3p/3ϫ PXRMRE plasmid were significantly (p Ͻ 0.01) lower than those of the control pGL3p plasmid (Fig. 2D). These activities were significantly (p Ͻ 0.01) restored by the transfection of AsO for miR-148a. These results underscore that miR-148a functionally recognizes PXRMRE148 to decrease the expression.
Effects of Overexpression or Inhibition of miR-148a on the PXR Protein Level in a Human Cell Line-We next examined the change in endogenous PXR protein expression by the overexpression or inhibition of miR-148a. By the transfection of the precursor for miR-148a into HepG2 cells that harbor the increased level of mature miR-148a, the PXR protein level was significantly (p Ͻ 0.05) decreased compared with the control (Fig. 3A). Conversely, by the transfection of the AsO for miR-148a into HepG2 cells, where the expression of mature miR-148a was extinguished, the PXR protein level was significantly (p Ͻ 0.05) increased compared with the control (Fig. 3B). Meanwhile, the expression level of RXR␣ protein, a heterodimer partner of PXR, was not affected by the overexpression or inhibition of miR-148a. It is well known that ligand-activated PXR activates the transcription of targets by binding to the responsive element. Using the pCYP3A4-362-7.7K plasmid containing the PXR-responsive element as a reporter construct, the changes in the PXR protein levels were monitored with the reporter activity (Fig. 3, C and D). The luciferase activity of pCYP3A4-362-7.7K plasmid was prominently (5.3-fold) increased by the treatment with rifampicin in HepG2 cells (Fig.   FIGURE 2. Repressive regulation of human PXR by miR-148a. A, complementarity of miR-148a to the predicted target sequence of human PXR. The potential miR-148a recognition element (PXRMRE148) is located on ϩ3359 to ϩ3386 in the 3Ј-UTR of human PXR mRNA, where the numbering refers to the 5Ј-end of mRNA as 1. B, the mature miR-148a levels in HepG2, HuH7, HLE, Caco-2, Caco-2/D (differentiated), LS180, HEK293, and MCF-7 cells were determined by real time RT-PCR analysis using an NCode miRNA first strand cDNA synthesis kit. The values were the mature miR-148a levels normalized with the U6 snRNA levels relative to that in HLE cells. C and D, luciferase assays were performed to investigate whether PXRMRE148 is functional in the regulation by miR-148a. The reporter constructs were transiently transfected with 4 pmol of the precursors for miR-148a or control into HEK293 cells (C) or 10 pmol of the AsO for miR-148a or control into HepG2 cells (D). The data were the firefly luciferase activities normalized with the Renilla luciferase activities relative to that of pGL3p plasmid. Each column represents the mean Ϯ S.D. of three independent experiments. **, p Ͻ 0.01, compared with pGL3p; † †, p Ͻ 0.01; † † †, p Ͻ 0.001, compared with AsO for control. ORF, open reading frame. 3E). Transfection of the precursor for miR-148a significantly decreased both the rifampicin-induced and basal transcriptional activities, resulting in a dose-dependent decrease of the induction. In contrast, the transfection of antisense for miR-148a significantly increased both the rifampicin-induced and basal transcriptional activity, resulting in a dose-dependent increase of the induction (Fig. 3F). These results suggest that miR-148a negatively regulates the expression of PXR protein and, subsequently, the induction of its targets.
Role of miR-148a-dependent PXR Regulation in the Induction of Endogenous CYP3A4 mRNA in a Human Cell Line-We next sought to examine whether the miR-148a-dependent change of the PXR protein level affects the CYP3A4 induction in human cells. LS-180 cells were used, because this cell line expressed relatively higher CYP3A4 mRNA than the other cell lines (data not shown). By the transfection of the precursor for miR-148a into the LS-180 cells, the PXR protein level was significantly (p Ͻ 0.01) decreased (Fig.  4A), concomitant with a dramatic increase of the mature miR-148a level, whereas the PXR mRNA level was not decreased at any time after the transfection (Fig. 4B). The RXR␣ protein level was not affected by the overexpression of miR-148a (Fig. 4A). As shown in Fig. 4C, the CYP3A4 mRNA level was significantly increased by the treatment with rifampicin (5.0-fold). However, this induction was diminished by the overexpression of miR-148a, although the basal CYP3A4 mRNA level was not affected. These results suggest that the miR-148a-dependent regulation of PXR affects the induction of CYP3A4.
The miR-148a-dependent PXR Regulation May Control CYP3A4 Expression in Human Liver Tissue-To further investigate the effects of the miR-148a-dependent regulation of PXR in human liver tissue, the relationships between the expression levels of miR-148a, PXR, and CYP3A4 were investigated using a panel of 25 human livers (supplemental Table 1). The expression levels of miR-148a were variable (95fold) in the panel of human livers. The miR-148a level in liver sample 18 was comparable with that in HepG2 cells. The PXR mRNA (75fold) and CYP3A4 mRNA (363-fold) were also variable. As shown in Fig. 1, the PXR mRNA level was not correlated with the PXR protein level. In contrast, the CYP3A4 mRNA level was significantly correlated (Rs ϭ 0.67, p Ͻ 0.001) with the CYP3A4 protein level (Fig. 5A). When the PXR protein/PXR mRNA ratios were calculated as an index of the translational efficiency of PXR, they were inversely correlated with the miR-148a level (Rs ϭ Ϫ0.41, p Ͻ 0.05) (Fig. 5B), suggesting that PXR is negatively regulated by miR-148a in human liver. The PXR protein level was significantly correlated with both the CYP3A4 mRNA level (Rs ϭ 0.47, p Ͻ 0.05) (Fig.  5C) and the CYP3A4 protein level (Rs ϭ 0.67, p Ͻ 0.001) (Fig.  5D). As summarized in Fig. 5E, the post-transcriptional regulation of PXR by miR-148a appeared to have substantial impact on the CYP3A4 level in human livers. The precursors for miR-148a or control (50 nM) (A) or AsOs for miR-148a or control (50 nM) (B) were transfected into HepG2 cells. After 72 h, the cells were harvested, and total RNA and nuclear extracts were isolated. A and B, the mature miR-148a levels were determined by real time RT-PCR analyses. The values were the mature miR-148a levels normalized with the U6 snRNA levels relative to control. The PXR and RXR␣ protein levels in nuclear extracts were determined by Western blot analyses. The values are the mean Ϯ S.D. for three independent experiments (*, p Ͻ 0.05, compared with control). C and D, schemes represent the principle of the reporter gene assay to evaluate the changes in the endogenous PXR protein level by the overexpression (C) or inhibition (D) of miR-148a. The pCYP3A4-362-7.7K plasmid contains the PXR-responsive elements, ER6 (Ϫ362 to ϩ11) and DR3 (Ϫ7836 to Ϫ7200), of the CYP3A4 gene upstream of the luciferase gene. E and F, the cells were transfected with the reporter plasmid and the precursors or AsOs. After 48 h, the cells were treated with 10 M rifampicin (squares) or 0.1% Me 2 SO (circles) for 24 h. The data were the firefly luciferase activities normalized with the Renilla luciferase activities relative to that of pGL3p plasmid without the precursors or AsOs. Each point represents the mean Ϯ S.D. of three independent experiments. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001, compared with control.

DISCUSSION
PXR regulates at least 40 genes encoding proteins responsible for the metabolism and elimination of drugs (14). The study of PXR regulation assists in the understanding of the inter-and intraindividual variability in the pharmacokinetics of drugs. Although many research groups have found variability in the PXR mRNA levels in human liver samples, the correlation with its protein level has not fully been investigated. In this study, we first demonstrated that there was no significant correlation between them in human livers. For human PXR, two splicing variants, including exon 1b (PXR.2) or deleting the 5Ј-end of the exon 5 (PXR.3), have been reported (24,25), which cannot be distinguished with our PCR primers. However, since the expression levels of these splicing variants were extremely low in our analysis (data not shown), consistent with a report by Lamba et al. (26), the dissociation of the PXR protein level with its mRNA level is not an artificial phenomenon. Identification of the miRNA recognition element in the human PXR gene suggested the involvement of miRNA in the regulation of PXR.
The luciferase assays showed that the endogenous and exogenous miR-148a negatively regulated the activity through PXRMRE148. In addition, the endogenous PXR protein level was diminished by the overexpression of miR-148a and ele-vated by its inhibition. These results clearly indicated that human PXR is post-transcriptionally regulated by miR-148a. The miR-148a-dependent changes of PXR protein affected the induction of CYP3A4 in LS180 cells. To further investigate whether the miR-148a affects the induction of other targets of PXR, we determined the expression levels of MDR1 and CYP2B6 in LS180 cells (data not shown). Rifampicin induced the MDR1 (5-fold) and CYP2B6 (2-fold) mRNAs, known targets of PXR (27,28), and the induction was attenuated by the overexpression of miR-148a. Thus, the miR-148a-dependent regulation of PXR appeared to affect its target genes in common.
Interestingly, the miR-148a recognition element is also present in the 3Ј-UTR of CYP3A4 mRNA. The complementarity of CYP3A4 with miR-148 (score 15.48, energy Ϫ24.27) was higher than that of human PXR (score 15.14, energy Ϫ19.52). To investigate whether CYP3A4 is directly regulated by miR-148a, luciferase assays were performed using a plasmid containing three copies of the miR-148a recognition element in the A and B, the precursors for miR-148a or control (50 nM) were transfected into LS180 cells. A, after 72 h, the cells were harvested, and total RNA and nuclear extracts were isolated. The mature miR-148a level was determined by real time RT-PCR analysis. The values are the mature miR-148a levels normalized with the U6 snRNA levels relative to control. The PXR and RXR␣ protein levels were determined by Western blot analysis. The values are the mean Ϯ S.D. of three independent experiments (**, p Ͻ 0.01, compared with control). B, after 24, 48, 72, and 96 h, the cells were harvested, and total RNA was isolated. The PXR mRNA level was determined by real time RT-PCR analysis. The values are the PXR mRNA levels normalized with the GAPDH mRNA levels relative to control. Each point represents the mean Ϯ S.D. of three independent experiments. C, precursortransfected LS180 cells were treated with 50 M rifampicin or 0.1% Me 2 SO for 24 h, and then total RNA was isolated. The CYP3A4 mRNA levels were determined by real time RT-PCR and normalized with the GAPDH mRNA level. The data are relative to that with the precursor for the control without rifampicin. Each column represents the mean Ϯ S.D. of three independent experiments. ***, p Ͻ 0.001; NS, not significant. FIGURE 5. Relationship between the expression levels of miR-148a, PXR, and CYP3A4 in human liver tissue. A, the CYP3A4 mRNA level was significantly correlated with the CYP3A4 protein levels. B, the miR-148a level was inversely correlated with the translational efficiency of PXR (PXR protein/mRNA ratio). C and D, the PXR protein level was significantly correlated with the CYP3A4 mRNA (C) and protein level (D). E, summary of the correlation analyses (*, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001).
CYP3A4 gene. However, unlike the PXRMRE, the element in CYP3A4 did not respond to miR-148a (supplemental Fig. 1), indicating that CYP3A4 is not directly regulated by miR-148a.
In the panel of human livers, the expression level of miR-148a was inversely correlated with the translational efficiency of PXR, supporting the role of miR-148a in the regulation of PXR in liver. The significant correlation between the CYP3A4 mRNA and the CYP3A4 protein level in human livers in this study, in accordance with previous studies (29,30), supported the finding that miRNA did not directly regulate the CYP3A4 expression. The PXR protein level was correlated with the CYP3A4 mRNA level in human liver, indicating that miR-148a affects the CYP3A4 expression through modulating PXR expression. The PXR protein level was not correlated with the CYP2B6 (Rs ϭ 0.31, p Ͼ 0.1) or MDR1 (Rs ϭ Ϫ0.20, p Ͼ 0.3) mRNA levels in the panel of human livers (data not shown), in contrast to CYP3A4. Thus, we speculate that the PXR does not largely affect the constitutive expression of CYP2B6 and MDR1 in the liver. In our panel of human livers, the CYP3A4 mRNA level was not correlated with the HNF4␣ protein level (Rs ϭ Ϫ0.14, p Ͼ 0.5) or CAR protein level (Rs ϭ 0.12, p Ͼ 0.5), indicating a significant contribution of PXR to the constitutive CYP3A4 level.
Most of the genes in the vertebrate nuclear receptor superfamily are strongly conserved between species. The ligandbinding domain of PXR shares amino acid identity of 75% between human and rodent, which can explain why the key ligands for PXR vary across species. Meanwhile, the DNA-binding domain of PXR shares more than 95% amino acid identity (1), inducing a similar set of genes. Most miRNAs are evolutionally conserved, which suggests that the miRNA-mediated regulation of certain genes would be common among species. The MRE148 is also identified in the 3Ј-UTR in mouse PXR (score 16.63, energy Ϫ21.4) and rat PXR (score and energy are not calculated at miRBase, but it has only a 1-base difference from the corresponding mouse sequence) at ϳ670 bp downstream of the stop codon, although the 3Ј-UTR of PXR is poorly conserved between humans and rodents. It is therefore possible that rodent PXR may also be regulated by miR-148a, suggesting that rodents might be useful model animals to investigate the role of miR-148a in drug metabolism and elimination in vivo.
In conclusion, we found that human PXR is post-transcriptionally regulated by miR-148a affecting the CYP3A4 level in human liver. This study would provide new insight into the unsolved mechanism of the large interindividual variability of CYP3A4 expression.