Transactivation of ABCG2 through a novel cis-element in the distal promoter by constitutive androstane receptor but not pregnane X receptor in human hepatocytes

A previous report demonstrated that treatment of human hepatocytes with phenobarbital, an activator of nuclear receptor constitutive androstane receptor (CAR), increases mRNA levels of an efflux transporter ABCG2, which is involved in the excretion of xenobiotics in liver and intestine. The results suggest that human CAR (hCAR) transactivates human ABCG2 (hABCG2) expression. In this study, we confirmed increase in ABCG2 mRNA levels in human hepatocytes after adenoviral expression of hCAR and treatment with its activator. Reporter assays suggested the existence of an hCAR-responsive element between -8000 and -7485 of hABCG2 promoter. Electrophoretic mobility shift assays and chromatin immunoprecipitation assays identified a DR5 motif (direct repeat separated by five nucleotides) within the region as a binding motif of hCAR/human retinoid X receptor α heterodimer. The introduction of mutations into the DR5 motif resulted in the complete loss of the hCAR-mediated transactivation. Interestingly, human pregnane X receptor, belonging to the same NR1I subfamily as CAR, did not activate any reporter gene containing the DR5 motif. Taken together, our present findings suggest that hCAR transactivates hABCG2 through the DR5 motif located in its distal promoter in human hepatocytes and that the motif prefers hCAR to pregnane X receptor.     

Human constitutive androstane receptor mediates induction of CYP2B6 gene expression by phenytoin

Compared with its rodent orthologs, little is known about the chemical specificity of human constitutive androstane receptor (hCAR) and its regulation of hepatic enzyme expression. Phenytoin (PHY), a widely used antiepileptic drug, is a potent inducer of CYP2B6 in primary human hepatocytes, but does not activate human pregnane X receptor (PXR) significantly in cell-based transfection assays at the same concentrations associated with potent induction of CYP2B6. Based on this observation, we hypothesized that PHY may be a selective activator of hCAR. In primary human hepatocytes, expression of CYP2B6 reporter genes containing phenobarbital-responsive enhancer module (PBREM) or PBREM/xenobiotic-responsive enhancer module (XREM) response elements were activated up to 14- and 28-fold, respectively, by 50 microm PHY. By contrast, parallel experiments in HepG2 cell lines co-transfected with an hPXR expression vector did not show increased reporter activity. These results indicated that a PXR-independent pathway, which is retained in primary hepatocytes, is responsible for PHY induction of CYP2B6. Further experiments revealed that PHY effectively translocates hCAR from the cytoplasm into the nucleus in both primary human hepatocytes and CAR(-/-) mice. Compared with vehicle controls, PHY administration significantly increased CYP2B6 reporter gene expression, when this reporter construct was delivered together with hCAR expression vector into CAR(-/-) mice. However, PHY did not increase reporter gene expression in CAR(-/-) mice in the absence of hCAR vector, implying that CAR is essential for mediating PHY induction of CYP2B6 gene expression. Taken together, these observations demonstrate that, in contrast to most of the known CYP2B6 inducers, PHY is a selective activator of CAR in humans.     

Expression of a human coxsackie/adenovirus receptor transgene permits adenovirus infection of primary lymphocytes

Replication-defective adenoviruses are effective vehicles for gene transfer, both for the repair of defective genes and for studies of gene function in primary cells. Many cell types, including lymphocytes, are refractory to adenovirus infection because they lack the Coxsackie/adenovirus receptor (CAR) needed for virus attachment. To extend the advantages of adenovirus-mediated gene transfer to primary lymphoid populations and other cell types lacking endogenous CAR, we produced a mouse that expresses human (h) CAR as a transgene under control of a murine MHC class I promoter. hCAR protein is expressed on T and B lymphocytes from a variety of organs (spleen, lymph node, bone marrow, thymus, and peritoneum). These lymphocytes are susceptible to adenovirus infection, as demonstrated by reporter green fluorescent protein gene expression, with the fraction of expressing cells as high as 70%. Some lymphocyte subpopulations required stimulation subsequent to adenovirus infection for reporter expression. This activation requirement is a restriction imposed by the promoter used in the adenovirus construct. In subpopulations requiring activation, the elongation factor 1 promoter was far superior to a hCMV promoter for directing green fluorescent protein production. We also find that hCAR mRNA is produced in nonlymphoid tissues from all founder lines, including tissues that do not express endogenous murine CAR, suggesting the opportunity for effecting gene delivery to and testing gene function in a wide variety of primary cell types previously resistant to gene transfer.     

Identification of a vitamin D responsive element in the promoter of the rat cytochrome P450(24) gene.

Mitochondrial cytochrome P450(24) expression in the vitamin D-degradation pathway is induced by 1,25-dihydroxyvitamin D3 [1,25-(OH)2D3]. The molecular basis of this enzyme regulation was investigated by isolating the rat P450(24) gene and examining the 5'-flanking region for possible cis-acting regulatory elements involved in the induction process. Constructs containing different lengths of 5'-flanking region of the gene were linked to a luciferase reporter gene and transiently co-transfected with a human vitamin D receptor (hVDR) expression vector (pRSV-hVDR) into COS-1 cells. These experiments showed that the flanking region from -298 to -122 directed a 24-fold increase in luciferase activity in response to 1,25-(OH)2D3 provided that the cells were co-transfected with pRSV-hVDR. Within this region, the sequence from position -171 to -123 conferred 1,25-(OH)2D3 responsiveness to both the native P450(24) promoter and the heterologous thymidine kinase promoter. Mutagenesis revealed that the sequence from position -150 to -136 is required for induction by 1,25-(OH)2D3 and that this sequence shares similarity to other vitamin D responsive elements (VDREs) reported for other genes. Gel shift mobility assays showed this region specifically bound a nuclear protein complex from 1,25-(OH)2D3 treated COS-1 cells that had been co-transfected with pRSV-hVDR. The retarded band was specifically competed with the well characterized VDRE from the mouse osteopontin gene. A VDRE at position -150 to -136 in the promoter of the rat P450(24) gene is identified in this study and found to be important in mediating the enhanced expression of the gene by 1,25-(OH)2D3.     

Biochemical evidence for a 170-kilodalton, AF-2-dependent vitamin D receptor/retinoid X receptor coactivator that is highly expressed in osteoblasts

Human vitamin D receptor (hVDR) fused to glutathione S-transferase was utilized to detect a VDR-interacting protein (VIP) of approximately 170 kDa. VIP(170) is expressed in osteoblast-like ROS 17/2.8 cells and, to a lesser extent, in COS-7 and HeLa cells. VIP(170) may be a coactivator because it interacts only with 1,25-dihydroxyvitamin D(3) (1,25(OH)(2)D(3)) ligand-bound hVDR and because a mutation (E420A) in the activation function-2 (AF-2) of hVDR abolishes both receptor-mediated transactivation and VIP(170) binding. Unlike L254G hVDR, a heterodimerization mutant with an intact AF-2, the E420A mutant is only partially attenuated in its association with the retinoid X receptor (RXR) DNA-binding partner. Finally, the ability of overexpressed hVDR to squelch glucocorticoid receptor-mediated transactivation is lost in both the L254G and E420A mutants. These results suggest that several protein-protein interactions, including VDR association with RXR and VIP(170), are required for stabilization of a multimeric complex that transduces the signal for 1,25(OH)(2)D(3)-elicited transactivation.     

Inhibition of apolipoprotein A-I gene expression by obesity-associated endocannabinoids

Obesity is associated with increased serum endocannabinoid (EC) levels and decreased high-density lipoprotein cholesterol (HDLc). Apolipoprotein A-I (apo A-I), the primary protein component of HDL is expressed primarily in the liver and small intestine. To determine whether ECs regulate apo A-I gene expression directly, the effect of the obesity-associated ECs anandamide and 2-arachidonylglycerol on apo A-I gene expression was examined in the hepatocyte cell line HepG2 and the intestinal cell line Caco-2. Apo A-I protein secretion was suppressed nearly 50% by anandamide and 2-arachidonoylglycerol in a dose-dependent manner in both cell lines. Anandamide treatment suppressed both apo A-I mRNA and apo A-I gene promoter activity in both cell lines. Studies using apo A-I promoter deletion constructs indicated that repression of apo A-I promoter activity by anandamide requires a previously identified nuclear receptor binding site designated as site A. Furthermore, anandamide-treatment inhibited protein-DNA complex formation with the site A probe. Exogenous over expression of cannabinoid receptor 1 (CBR1) in HepG2 cells suppressed apo A-I promoter activity, while in Caco-2 cells, exogenous expression of both CBR1 and CBR2 could repress apo A-I promoter activity. The suppressive effect of anandamide on apo A-I promoter activity in Hep G2 cells could be inhibited by CBR1 antagonist AM251 but not by AM630, a selective and potent CBR2 inhibitor. These results indicate that ECs directly suppress apo A-I gene expression in both hepatocytes and intestinal cells, contributing to the decrease in serum HDLc in obese individuals.     

Analysis of the 5-lipoxygenase promoter and characterization of a vitamin D receptor binding site

1,25-Dihydroxyvitamin D(3) (1,25(OH)(2)D(3)) and transforming growth factor beta (TGFbeta) potently induce 5-lipoxygenase (5-LO) in myeloid cells. We analyzed vitamin D receptor (VDR) binding to putative vitamin D response elements within the 5-LO promoter and analyzed its function by reporter gene analysis. Binding of VDR and retinoid X receptor to the promoter region was shown in DNase I footprinting, electrophoretic mobility shift and chromatin immunoprecipitation assays. However, the identified VDR binding region did not mediate induction of reporter gene activity by 1,25(OH)(2)D(3)/TGFbeta, neither in the 5-LO promoter context nor with the thymidine kinase (tk) promoter. Insertion of the rat atrial natriuretic factor VDRE in reporter plasmids containing the 5-LO promoter diminished induction by 1,25(OH)(2)D(3)/TGFbeta as compared with the tk promoter. Similarly, low inductions were obtained when cells were transiently or stably transfected with constructs containing various 5-LO promoter regions. Concerning basal promoter activity, we identified a positive regulatory region (-779 to -229), which includes the VDR binding region, in 5-LO-positive MonoMac6 cells. In summary, the VDR/RXR complex binds to putative VDREs in the 5-LO promoter, but other sequences outside the 5-LO promoter seem to be responsible or additionally required for the prominent induction of 5-LO mRNA expression by 1,25(OH)(2)D(3) and TGFbeta.     

Overexpression of the human vitamin D3 receptor in mammalian cells using recombinant adenovirus vectors.

The human vitamin D3 receptor (hVDR) cDNA was cloned into the E1 region of the adenovirus genome to generate recombinant viruses which were used to infect 293 (adenovirus-transformed human fetal kidney) cells. High salt extracts from cells infected with the recombinant viruses were subjected to immunoblot analysis using a monoclonal antibody to chicken VDR and were shown to contain large quantities of a protein of approximately 50 kDa with a migration identical to that of the hVDR in T47D (human mammary adenocarcinoma) cells. Scatchard analysis showed that the infected cells express approximately 100-fold more receptor than T47D cells and that this receptor binds 1,25-dihydroxyvitamin D3 with high affinity. The overexpressed hVDR also binds to DNA-cellulose and is eluted with a KCl concentration similar to that determined for fully active endogenous VDR. Nuclear extracts from cells infected with the hVDR-expressing adenoviruses contain an activity that specifically binds an oligonucleotide with sequences from the rat osteocalcin vitamin D3 response element, as determined by gel mobility shift. This interaction can be inhibited by the presence of an anti-VDR antibody, but not by nonspecific immunoglobulins. We conclude, therefore, that the overexpressed receptor has the ligand- and DNA-binding characteristics defined for endogenous VDR and that adenoviruses can be used to efficiently express large quantities of functional hVDR in a human cell line. Finally, a second binding activity, specific for the vitamin D response element, but distinct from the VDR, has been identified in extracts from uninfected cells.     

Exons and functional regions of the human vitamin D receptor gene around and within the main 1a promoter are well conserved among mammals

The human vitamin D receptor (hVDR) gene encompasses eight exons (2–9) in the so-called coding region and six more exons (1a–1f) in the so-called regulatory region, which contains several reported promoters. Evolutionary comparison performed on the VDR promoter sequences of a dozen of mammalian species shows a very high conservation of numerous regions around and in the 1a promoter, including exons 1e, 1a and 1d, and the Sp1 site region. This suggests that the so-called 1a promoter is well conserved among mammals. Homology among mammals also concerns three functional SNP site regions of the hVDR 1a promoter, the 1e-G-1739A SNP region (a Cdx-2 binding site), and both 1a-G-1521C and 1a-A-1012G sites, the 1a-1012A being located within a GATA site. Interestingly, the 1521G and 1012A nucleotides are being evolutionary conserved, suggesting that the 1521C/1012G haplotype, which is found in human chromosomes (43% of Caucasians), is a human specificity.     

Gangliosides act as co-receptors for Salmonella enteritidis FliC and promote FliC induction of human beta-defensin-2 expression in Caco-2 cells

Antimicrobial peptides such as defensins are crucial for host defense at mucosal surfaces. We reported previously that Salmonella enteritidis flagellin (FliC) induced human beta-defensin-2 (hBD-2) mRNA expression in Caco-2 cells via NF-kappaB activation (Ogushi, K., Wada, A., Niidome, T., Mori, N., Oishi, K., Nagatake, T., Takahashi, A., Asakura, H., Makino, S., Hojo, H., Nakahara, Y., Ohsaki, M., Hatakeyama, T., Aoyagi, H., Kurazono, H., Moss, J., and Hirayama, T. (2001) J. Biol. Chem. 276, 30521-30526). In this study, we examined the role of ganglioside as co-receptors with Toll-like receptor 5 (TLR5) on FliC induction of hBD-2 expression in Caco-2 cells. Exogenous gangliosides suppressed FliC induction of hBD-2 promoter activity and binding of FliC to Caco-2 cells. Incorporation of exogenous ganglioside GD1a into Caco-2 cell membranes increased the effect of FliC on hBD-2 promoter activity. In support of a role for endogenous gangliosides, incubation of Caco-2 cells with dl-threo-2-hexadecanoylamino-3-morpholino-1-phenylpropanol, a glucosylceramide synthase inhibitor, reduced FliC induction of hBD-2 promoter activity. GD1a-loaded CHO-K1-expressing TLR5 cells had a higher potential for hBD-2 induction following FliC stimulation than GD1a-loaded CHO-K1 cells not expressing TLR5. FliC increased phosphorylation of mitogen-activated protein kinase, p38, and ERK1/2. Exogenous gangliosides GD1a, GD1b, and GT1b each suppressed FliC induction of p38 and ERK1/2 phosphorylation. Furthermore, FliC did not enhance luciferase activity in Caco-2 cells transfected with a plasmid containing a mutated activator protein 1-binding site. These results suggest that gangliosides act as co-receptors with TLR5 for FliC and promote hBD-2 expression via mitogen-activated protein kinase.     

Molecular mechanism of the vitamin D antagonistic actions of (23S)-25-dehydro-1alpha-hydroxyvitamin D3-26,23-lactone depends on the primary structure of the carboxyl-terminal region of the vitamin d receptor

We reported that (23S)-25-dehydro-1alpha-hydroxyvitamin D(3)-26,23-lactone (TEI-9647) antagonizes vitamin D receptor (VDR)-mediated genomic actions of 1alpha,25-dihydroxyvitamin D(3) [1alpha,25(OH)(2)D(3)] in human cells but is agonistic in rodent cells. Human and rat VDR ligand-binding domains are similar, but differences in the C-terminal region are important for ligand binding and transactivation and might determine the agonistic/antagonistic effects of TEI-9647. We tested TEI-9647 on 1alpha,25(OH)(2)D(3) transactivation using SaOS-2 cells (human osteosarcoma) or ROS 24/1 cells (rat osteosarcoma) cotransfected with human or rodent VDR and a reporter. In both cell lines, TEI-9647 was antagonistic with wild-type human (h)VDR, but agonistic with overexpressed wild-type rat (r)VDR. VDR chimeras substituting the hVDR C-terminal region (activation function 2 domain) with corresponding rVDR residues diminished antagonism and increased agonism of TEI-9647. However, substitution of 25 C-terminal rVDR residues with corresponding hVDR residues diminished agonism and increased antagonism of TEI-9647. hVDR mutants (C403S, C410N) demonstrated that Cys403 and/or 410 was necessary for TEI-9647 antagonism of 1alpha,25(OH)(2)D(3) transactivation. These results suggest that species specificity of VDR, especially in the C-terminal region, determines the agonistic/antagonistic effects of TEI-9647 that determine, in part, VDR interactions with coactivators and emphasize the critical interaction between TEI-9647 and the two C-terminal hVDR Cys residues to mediate the antagonistic effect of TEI-9647.     

Dual effect of dexamethasone on CYP3A4 gene expression in human hepatocytes. Sequential role of glucocorticoid receptor and pregnane X receptor.

Although CYP3A induction by dexamethasone has been extensively documented, its mechanism is still unclear because both the role of the glucocorticoid receptor and the ability of dexamethasone to activate the human pregnane X receptor have been questioned. In an attempt to resolve this problem, we investigated the response of CYP3A4 to dexamethasone (10 nm-100 microm) in primary human hepatocytes and HepG2 cells, using a variety of methods: kinetic analysis of CYP3A4 and tyrosine aminotransferase expression, effects of RU486 and cycloheximide, ligand binding assay, cotransfection of HepG2 cells with CYP3A4 reporter gene constructs and vectors expressing the glucocorticoid receptor, pregnane X receptor or constitutively activated receptor. In contrast to rifampicin (monophasic induction), dexamethasone produces a biphasic induction of CYP3A4 mRNA consisting of a low-dexamethasone component (nmol concentrations) of low amplitude (factor of 3-4) followed by a high-dexamethasone component (supramicromolar concentrations) of high amplitude (factor of 15-30). We show that the low-dexamethasone component results from the glucocorticoid receptor-mediated expression of pregnane X receptor and/or constitutively activated receptor which, in turn, are able to transactivate CYP3A4 in a xenobiotic-independent manner. At supramicromolar concentrations (>10 microm), dexamethasone binds to and activates pregnane X receptor thus producing the high-dexamethasone component of CYP3A4 induction. We conclude that, in contrast to the other xenobiotic inducers of CYP3A4, glucocorticoids play a dual role in CYP3A4 expression, first by controlling the expression of PXR and CAR under physiological conditions (submicromolar concentrations) through the classical glucocorticoid receptor pathway, and second by activating the pregnane X receptor under bolus or stress conditions (supramicromolar concentrations).