Interleukin-1β represses MRP2 gene expression through inactivation of interferon regulatory factor 3 in HepG2 cells



The human multidrug resistance protein 2 (MRP2/ABCC2), expressed on the bile canalicular membrane, mediates the multispecific efflux of several organic anions, including conjugates of glucuronate, sulfate, and glutathione. Expression of MRP2 can be altered in response to environmental stimuli such as cholestasis and jaundice. We previously reported that MRP2 mRNA expression levels are decreased in the nontumorous part of hepatitis C virus-infected human liver tissues, and that inflammatory cytokines inhibit MRP2 expression in human hepatic (HepG2) cells. We investigated the molecular mechanisms by which inflammatory cytokines modulate MRP2 gene expression in hepatic cells. Treatment of human hepatic cells with interleukin-1β (IL-1β) or tumor necrosis factor α resulted in a decrease in the protein and mRNA levels of MRP2. IL-1β inhibited the transcriptional activity of MRP2 promoter constructs by 40%, and this inhibition of MRP2 promoter activity was mediated through the interferon stimulatory response element (ISRE). Electrophoretic mobility shift assays with IL-1β-treated nuclear extracts showed a decrease in the formation of DNA protein complexes, specifically those including interferon regulatory factor 3 (IRF3). Expression of recombinant human IRF3 increased MRP2 promoter activity. Treatment with a specific extracellular signal-regulated kinase inhibitor relieved IL-1β-induced MRP2 mRNA downregulation and abrogated the binding of IRF3 to the ISRE element. In conclusion, IL-1β induces downregulation of the MRP2 gene by inactivating IRF3 binding to ISRE on the MRP2 promoter in human hepatic cells; this inactivation is accomplished via interference with the extracellular signal-regulated kinase pathway. (HEPATOLOGY 2004;39:1574–1582.)

Transport proteins in the basolateral and canalicular membranes of hepatocytes mediate transport of organic solutes into and from the liver. Biliary elimination of both endogenous compounds and exogenous drugs or poisons is a major physiological self-defense role of various transporters.1 One of the major transport systems of this type involves the adenosine triphosphate binding cassette (ABC) transporter family. Among the numerous ABC transporters, P-glycoprotein (P-gp/ABCB1) and multidrug resistance protein 2 (MRP2/ABCC2) have been shown to mediate transport of cationic and anionic compounds into bile, and bile salt exporting pump (BSEP/ABCB11) has been shown to export bile acids.2, 3MRP2 initially was identified as a multispecific organic anion transporter in the canalicular membrane of hepatocytes and was shown to transport a wide range of conjugated compounds, including leukotriene C4, glucuronidase and sulfated molecules, and molecules conjugated to glutathione.4–6 Inherited defects in MRP2 result in Dubin-Johnson syndrome, a congenital disease associated with chronic hyperbilirubinemia and jaundice.7–9 Liver diseases are characterized by disturbances in the hepatobiliary transport of endogenous and exogenous compounds. Hepatic expression of ABC transporters can be influenced by a number of endogenous and environmental factors. Inflammation is induced under various pathological conditions, such as carcinogenesis, cholestasis, and regeneration by partial hepatectomy and by lipopolysaccharides.10–14 Administrations of inflammatory cytokines interleukin (IL)-6 and IL-1β has been shown to result in a 20% to 60% decrease in the mRNA levels of MRP2, organic anion-transporting polypeptide 1, organic anion-transporting polypeptide 2, and bile salt exporting pump in the mouse liver compared with untreated controls.15 The decreased hepatic expression of these transporters affects the cellular efflux of various physiological compounds, suggesting that these cytokines may play a key role in the hepatic expression of anion transporters in inflammatory cholestasis.15 Altered expression of the organic anion-transporting polypeptide and Mrp transporters also is expected to influence the pharmacokinetics of various drugs that are transported by these proteins.

The transcription factor interferon regulatory factor 3 (IRF3) is expressed constitutively in many cell types, and its expression is thought to contribute to self-defense from viral infection by inducing type I interferons.16, 17 IRF3 is required for the expression of interferon β and the chemokine regulated on activation, normal T cell expressed and secreted (RANTES) in response to viral infection. In unstimulated cells, IRF3 is present dominantly in the cytoplasm and is phosphorylated in the N terminal domain. After viral infection, the C terminal domain of IRF3 is phosphorylated, leading to dimerization and interaction with the coactivator CBP/p300.18 This complex is then translocated to the nucleus, where it activates a promoter containing the IRF3 binding site.

We previously compared the expression of ABC transporters between the hepatitis C virus (HCV)-infected and non-HCV-infected liver in patients with hepatic cancers and reported that the expression levels of MRP2 mRNA and MRP2 protein were decreased significantly and specifically in the nontumorous part of HCV-infected liver tissue.19 However, the expression levels of other ABC transporters—MDR1, MDR3, MRP1, and MRP3—were not significantly different between the HCV-infected and non-HCV-infected liver.19 We hypothesized that the downregulation of MRP2 mRNA levels after cytokine treatment occurred primarily at the transcriptional level, and that this reduction in MRP2 transcription could be the result of alteration in regulatory nuclear transcription factors. We previously characterized regulatory elements in the human MRP2 promoter in liver cells.20 In the present study, we show that MRP2 mRNA transcription decreased 24 hours after IL-1β and tumor necrosis factor α (TNFα) treatment. Nuclear extract treated with IL-1β showed a marked decrease in DNA protein complexes, including IRF3 complexes, and nuclear accumulation of IRF3 was decreased by treatment with IL-1β. The IL-1β−induced MRP2 mRNA downregulation was relieved by a specific extracellular signal-regulated kinase (ERK) inhibitor. We hypothesize that this inflammatory cytokine-induced downregulation of MRP2 is the result of reductions in both the nuclear accumulation of IRF3 and the binding of IRF3 to interferon stimulatory response element (ISRE), and that these reductions occur via the ERK pathway.


ABC transporter, adenosine triphosphate binding cassette transporter; P-gp, P-glycoprotein; MRP2, multidrug resistance protein 2; IL, interleukin; IRF3, interferon regulatory factor 3; RANTES, regulated on activation, normal T cell expressed and secreted; HCV, hepatitis C virus; TNFα, tumor necrosis factor α; ERK, extracellular signal-regulated kinase; ISRE, interferon stimulatory response element; PCR, polymerase chain reaction; C/EBP, CCAAT enhancer-binding protein; HNF, hepatocyte nuclear factor; USF, upstream stimulatory factor; EMSA, electrophoretic mobility shift assay; NE, nuclear extract; RXR, retinoid X receptor.

Materials and Methods

Cell Culture.

Human hepatoblastoma HepG2 cells were cultured in Dulbecco's modified Eagle's medium (Nissui Seiyaku, Tokyo, Japan) containing 10% fetal calf serum.19, 20 Normal human hepatocytes hNHeps (Sanko Junyaku, Japan) were cultured according to the manufacture's protocol.

Antibodies and Drugs.

The antibodies used in these experiments are as follows: MRP2 (M2III-6; Alexis, San Diego, CA), P-gp (C219; Centacor, Malvern, PA), IRF3 (FL425; Santa Cruz Biotech, Santa Cruz, CA), high mobility group protein-I (T-16; Santa Cruz Biotech), and glucose-6-phosphate dehydrogenase (A9521; Sigma-Aldrich, Saint Louis, MO). All IRFs and Sp1 antibodies were purchased from Santa Cruz Biotech. The drugs Actinomycin D, PD98059, SB203580, and SP600125 were purchased from Carbiochem (San Diego, CA).

Preparation of Protein.

For whole-cell lysate, cells were washed with ice cold phosphate-buffered saline and solubilized in 0.5 mL of radio-immunoprecipitation assay buffer (50 mM Tris HCl, 150 mM NaCl, 1% Triton X-100, 0.1% SDS, 0.5% deoxycholate, 1 mM phenylmethyl sulfonyl fluoride, and 1 mg/mL trypsin inhibitor) for 30 minutes on ice, then centrifuged at 15,000g for 10 minutes at 4°C. For nuclear and cytoplasmic extract, HepG2 and hNHeps were harvested by exposure to trypsin, resuspended in 200 μL of an ice-cold solution containing 10 mM HEPES NaOH (pH 7.9), 10 mM KCl, 0.2 mM ethylenediaminetetraacetic acid (EDTA), 0.2 mM ethylene glycol bis(βaminoethyl ether) (EGTA), 0.5 mM dithiothreitol, and 0.5 mM phenylmethylsulfonyl fluoride, and incubated on ice for 15 minutes. The cells were then lysed by passing 10 times through a 25-gauge needle attached to a 1-mL syringe, and the lysate was centrifuged for 40 seconds in a microcentrifuge. The supernatant was stored for cytoplasmic extract. The nuclear pellet was resuspended in 100 μL of an ice-cold solution containing 20 mM HEPES NaOH (pH 7.9), 0.4 M NaCl, 0.75 mM spermidine, 0.15 mM spermine, 0.2 mM EDTA, 0.2 mM EGTA, 0.5 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, and 25% (vol/vol) glycerol, incubated for 30 minutes on ice with frequent gentle mixing, and then centrifuged for 20 minutes at 4°C in a microcentrifuge to remove insoluble material. The resulting supernatant (nuclear extract) was stored at −80°C.

Northern Blot Analysis.

Total RNA was isolated using RNeasy spin columns (Qiagen, Hilden, Germany). 10 μg of total RNA from HepG2 cells was separated on a 1% formaldehyde-agarose gel and transferred to a membrane as described previously.19 An MRP2 complementary DNA (cDNA) probe (−28 to 513) was synthesized by polymerase chain reaction (PCR).

Reporter Gene Vector Constructs.

The fragments of the 5′ region of the MRP2 gene were ligated into the SacI and HindIII sites of pGL3-basic (Promega, Madison, WI). All plasmids were analyzed by restriction enzyme digestion, and the promoter inserts were sequenced. Site-directed mutagenesis of ISRE in p-491 MRP2 Luci was performed by a PCR-based method. The promoter sequence was amplified with Taq polymerase, a 5′-primer 491TAGGAGCTCTAGCGACTGATGCCAC and a 3′-primer that introduces a specific mutation into the ISRE region (5′-AGAAGCGAAACT-3′ to 5′-AcgtGCGcgtCT-3′). Amplification was also performed with a 5′-primer that introduces second specific mutation into the ISRE region and a 3′-primer AAGCTTGATTCCTGGACTGCGTC. Mutant constructs of CCAAT enhancer-binding protein β (C/EBPβ), hepatocyte nuclear factor 1 (HNF1), and upstream stimulatory factor (USF) were made using the same method (for C/EBPβ: 5′primer GAACTTTTAACCGCCTGTATTATG; 3′ primer AAGCTTGATTCCTGGACTGCGTC; for HNF1: 5′ primer GGCAAGGTCGGCGATTAAATGG; 3′ primer CCATTTAATCGCCGACCTTGCC; and for USF 5′ primer GGCTTTTTAGTTGTATGTCCATCC; 3′primer GGATGGACATTGTACTAAAAAGCC). A second PCR was then performed with Taq polymerase using the first PCR products as a template. The PCR product was cloned into pGEM-Teasy vector, which was subsequently digested with SacI and HindIII fragments. The fragments were ligated into the SacI and HindIII sites of pGL3-basic (Promega).

Luciferase Assay.

HepG2 cells were transfected by the Lipofectamine method as previously described.21 Briefly, a mixture of 5 μg of Lipofectamine 2000 (Life Technologies, Grand Island, NY), and 1 μg of reporter plasmid was incubated with the cells for 6 hours. Then, fresh medium containing 20 ng/mL IL-1β was added, and the cells were cultured for an additional 24 hours. In separate experiments, 20 μM PD98059 was administered for 30 minutes before IL-1β stimulation. 100 ng of pRL-TK (Promega) was cotransfected as a control for transfection. Luciferase activity was measured using a dual luciferase assay system (Promega). Twenty, 100, or 200 ng of Flag-only vector or Flag-IRF3 vector was cotransfected to investigate the effect of IRF3 function on the MRP2 promoter. Promoter activities are given as the mean ± SD of triplicate transfections. The level of significance of promoter activities in the presence of regular substrates was determined using Student's t test.

Electrophoretic Mobility Shift Assay (EMSA).

Nuclear extracts were incubated for 30 minutes on ice in a final volume of 20 μL of reaction mixture containing 25 mM HEPES (pH 7.5), 100 mM KCl, 1 mM EDTA, 10% glycerol, 0.7 mM DTT, 10 ng of poly (dIdC), and 1 × 104 cpm of 32P-labeled oligonucleotide probe in the absence or presence of wild-type or mutant (mt) competitors. The samples were electrophoresed on an 8% polyacrylamide gel (polyacrylamide/bisacrylamide ratio, 80:1) in Tris borate buffer. The DNA sequence of the sense strand of the wild-type MRP2ISRE oligonucleotide is GCAGCAGAAGCGAAACTGCAC, and that of the mutant MRP2ISRE is GCAGCAcgtGCGcgtCTGCAC. For supershift experiments, 2 μg of each antibody was added to the reaction mixture.


IL-1β and TNFα Reduced the Expression of the MRP2 Protein and mRNA.

We examined whether the MRP2 protein levels were affected by treatment of IL-1β and TNFα in HepG2 cells. Treatment with various doses of IL-1β and TNFα up to 20 ng/mL for 24 hours markedly reduced the MRP2 protein level in a dose-dependent manner, although P-gp levels remained the same. (Fig. 1A). We next examined the effect of IL-1β and TNFα stimulation on MRP2 mRNA expression levels. MRP2 mRNA levels were markedly reduced in a dose- and time- dependent manner (Fig. 1B,C). Treatment with IL-1β at 10 ng/mL or TNFα at 1 ng/mL for 24 hours reduced cellular MRP2 mRNA levels to 40% or less of the levels in the untreated control (Fig. 1B). Treatment for 12 hours or longer with 20 ng/mL IL-1β or TNFα decreased MRP2 mRNA levels to 30% to 40% of the levels in controls at time 0 (Fig. 1C). We next examined whether IL-1β signaling inhibitors could affect the IL-1β-induced downregulation of MRP2 gene expression. IL-1β is known to regulate several signal transduction cascades, including the three main kinase cascades, mitogen-activated protein kinase-ERK, c-Jun NH2-terminal kinase (JNK), and p38 mitogen-activated protein kinase.22, 23 Coadministration of SB203580 (p38 mitogen-activated protein kinase inhibitor) and SP600125 (JNK inhibitor) with IL-1β had no effect on the IL-1β-mediated downregulation of MRP2 mRNA levels (Fig. 2, lanes 2, 3, and 5), whereas PD98059 (ERK1/2 inhibitor) almost completely relieved the downregulatory effect of IL-1β (Fig. 2, lanes 2 and 4). There seemed to be no inhibitory effect of PD98059 alone on MRP2 mRNA expression (Fig. 2, lane 7). These results suggest that MRP2 mRNA downregulation by IL-1β involves the ERK pathway.

Figure 1.

Alteration of multidrug resistance protein 2 (MRP2) and P-glycoprotein (P-gp)/MDR1 protein and mRNA levels in HepG2 in response to interleukin-1β (IL-1β) or tumor necrosis factor-α (TNFα) administration. (A) For Western blot analysis, 200 μg of protein was loaded, and blots were incubated with an antibody, M2III-6 for MRP2 and C219 for P-gp/MDR1. (B,C) Northern blot analysis. (B) HepG2 cells were treated with the indicated dose of IL-1β (left) or TNFα (right) for 24 hours. (C) HepG2 cells were treated with 20 ng/mL IL-1β (left) or TNFα (right) for the indicated time. Ten micrograms total RNA was loaded and hybridized with MRP2 cDNA. The results were representative of three experiments.

Figure 2.

Extracellular signal-regulated kinase (ERK) pathways were involved in interleukin-1β (IL-1β)-induced MRP2 downregulation. 20 μM PD98059 (MEK1/2 inhibitor), 25 μM SB203580 (p38 inhibitor), or 20 μM SP600125 (JNK inhibitor) were pretreated for 30 minutes before IL-1β stimulation (20 ng/mL) in HepG2 cells. After 24 hours, total RNA were harvested and hybridized with multidrug resistance protein 2 (MRP2) cDNA. Relative MRP2 mRNA expression levels are shown, compared with the expression in untreated cells. This result was representative of three experiments.

To understand how MRP2 mRNA expression was downregulated by IL-1β, we initially examined whether the stability of MRP2 mRNA was altered. The MRP2 mRNA stability was examined, in the absence or presence of IL-1β, by blocking synthesis with 5 μg/mL of actinomycin D (Fig. 3). We observed that MRP2 mRNA was degraded at similar half-lives of approximately 12 hours under both conditions (Fig. 3), suggesting that IL-1β-induced downregulation of MRP2 mRNA was not the result of destabilization of mRNA by IL-1β.

Figure 3.

Rates of MRP2 mRNA decay. 5 μg/mL actinomycin D was incubated for 30 minutes with HepG2 cells, and interleukin-1β (IL-1β; 20 ng/mL) was administrated for the indicated time. Total RNA was harvested, and 10 μg of RNA was loaded on a 1% formaldehyde-agarose gel and transferred to a membrane. This result was representative of three experiments.

Human MRP2 Gene Promoter Activity was Inhibited by IL-1β Through an Interferon Stimulatory Response Element at −179/−146bp.

We investigated human MRP2 promoter activity in response to IL-1β administration using a series of 5′-promoter deletion analysis. Deleted promoter fragments were ligated into the reporter gene vector pGL3 basic and were transiently transfected into HepG2 cells. The luciferase activity of the complete series of 5′-deleted MRP2 promoter constructs is shown in Fig. 4. Compared with p-1659, the luciferase activity decreased to 30% when p-491 was assayed, suggesting that a putative positive cis-element is localized in the −1659/−491 bp region. We also observed an approximately 50% increase in luciferase activity by p-179 as compared with p-491, suggesting that a negative regulatory element is localized in the −491/−179 bp region. Administration of IL-1β reduced MRP2 promoter activity by 30% to 50% when p-2653, p-1659, p-491, and p-179 were assayed (Fig. 4). However, IL-1β failed to reduce the MRP2 promoter activity when p-146 and p-21 constructs were assayed. These results suggested that a putative IL-1β response element is located on the −179/−146 region. We identified an ISRE within the −179/−146 region (Fig. 4). To examine whether mutations within the ISRE binding motif affected the downregulation of the MRP2 promoter activity in response to IL-1β, we performed a mutagenesis analysis of the promoter. The construct designed contained a specific mutation in the consensus sequence of ISRE on the MRP2 promoter (AGAAGCGAAACT to AcgtGCGcgtCT). To check the adjacent transcription factors, we also made C/EBPβ, HNF 1, and USF mutant constructs (Fig. 5). We transfected the mutant p-491MRP2-Luci constructs into HepG2 cells in the absence or presence of IL-1β. Compared with the wild-type construct, the ISRE mutant reporter gene constructs showed a slightly lower basal activity in comparison with the wild type. Introduction of mutation into USF (USFmt) did not affect the basal promoter activity, whereas the HNFmt construct showed a marked decrease in the basal promoter activity, suggesting that HNF plays a key role in the MRP2 basal promoter activity. C/EBPmt construct enhanced the basal promoter activity by approximately 170% compared with the wild-type construct, indicating a negative regulatory role of C/EBPβ. When the p-491 MRP2 ISREmt was treated with IL-1β, no suppression of reporter gene activity was observed. There was a significant decrease in the promoter activity of IL-1β-treated USFmt, HNFmt, and C/EBPβmt reporter constructs (Fig. 5). These results indicated that the ISRE element specifically contributes to both basal promoter activity and IL-1β responsive downregulation in MRP2 promoter constructs.

Figure 4.

Maps of the MRP2 promoter luciferase constructs and the transcriptional activity of multidrug resistance protein 2 (MRP2) promoter-luciferase fusion plasmids transfected into HepG2 cells. Various lengths of human MRP2 gene promoter plasmids containing the luciferase gene in the upstream region were constructed. Putative transcription factor binding sites on the MRP2 promoter are also indicated. Luciferase activities were measured after 24 hours and were correlated for differences in transfection efficiency by Renilla luciferase activity, then normalized to the activity of the p-491 MRP2-Luci construct transfected into cells. Data are shown as the mean ± SD (error bars) of three independent experiments. *P < .01. ISRE, interferon stimulatory response element; USF, upstream stimulatory factor; HNF-1, hepatic nuclear factor 1.

Figure 5.

Transcriptional activity of p-491 MRP2-Luci and 4 p-491 MRP2 mutant-Luci (interferon stimulatory response element [ISRE], C/EBPβ, hepatic nuclear factor 1 [HNF1], and upstream stimulatory factor [USF]) transfected into HepG2 cells with or without 20 ng/mL of interleukin-1β (IL-1β) for 24 hours. Various mutant forms of human MRP2 gene promoter plasmids containing the luciferase gene in the upstream region were constructed. Luciferase activities were measured after 24 hours and were corrected for differences in transfection efficiency by Renilla luciferase activity, then normalized to the activity of the p-491 MRP2-Luci construct transfected into cells. This result was representative of three experiments. *P < .01. MRP2, multidrug resistance protein 2.

To investigate whether an ERK inhibitor, PD98059, affects IL-1β-mediated suppression of MRP2 promoter activity, we performed a transient transfection assay with p-491 MRP2-Luci and p-491(ISREmt) MRP2-Luci constructs in the presence or absence of this drug (Fig. 6). IL-1β again reduced the promoter activity of p-491 MRP2, and pretreatment with PD98059 significantly reduced IL-1β-mediated suppression of MRP2 promoter activity. The basal promoter activity of p-491 (ISREmt) MRP2 was decreased relative to that of p-491 MRP2, and the suppression of the promoter activity of the mutant construct by IL-1β was not significant (Fig. 6). Coadministration of PD98059 seemed to have no effect on the p-491MRP2 (ISREmt)-driven luciferase activity.

Figure 6.

Effects of PD98059 on interleukin-1β (IL-1β)-induced MRP2 promoter suppression. p-491MRP2-Luci and p-491MRP2(ISREmt)-Luci were transfected into HepG2 cells. Six hours later, 20 μM PD98059 was administrated for 30 minutes and IL-1β was treated (20 ng/mL; 24 hours). This result was representative of three experiments. *P < .05. MRP2, multidrug resistance protein 2; ISRE, interferon stimulatory response element.

Binding of IRF3 to ISRE on MRP2 Promoter is Reduced by IL-1β.

EMSAs were performed to investigate the binding of IRF3 to MRP2 promoter ISRE elements in nuclear extract from HepG2 cells. EMSA using radiolabeled ISRE oligonucleotides and nuclear extract showed two bands corresponding to specifically bound complexes (Fig. 7A). These ISRE–protein complex bands were reduced by treatment with 20 ng/mL IL-1β for 6 to 8 hours in the absence of an ERK inhibitor (Fig. 7A). By contrast, no decrease of ISRE–protein complex formation was observed when PD98059 was coadministered with IL-1β (Fig. 7B).

Figure 7.

Binding of IRF3 to MRP2 promoter interferon stimulatory response elements (ISREs). Electrophoretic mobility shift assay (EMSA) was performed using radiolabeled probes, including interferon stimulatory response element (ISRE; −171/−159, 5′-GCAGCAGAAGCGAAACTGCAC-3′). For the competition assay, the unlabeled ISRE mutant probe (−179/−153: 5′-GCAGCAcgtGCGcgtCTGCA-3′) was used as a competitor. (A) Effects of interleukin-1β (IL-1β) on nuclear protein-ISRE binding. Cells were incubated with 20 ng/mL IL-1β for 0 to 8 hours. The nuclear extracts were prepared, and the extracts (4 μg of protein) that were incubated with 32P-labeled oligonucleotide were resolved by gel electrophoresis. (B) Effects of PD98059 on IL-1β-induced MRP2 downregulation. Before IL-1β stimulation, 20 μM PD98059 was administrated for 30 minutes. (C) Super shift and competition assay using nuclear extract with or without IL-1β for 8 hours. Anti-IRF1-4, 7-9, Sp1, and RXRα antibodies were added to the nuclear extracts. A 50-fold excess of the unlabeled oligonucleotide was added for the competition. (D) EMSA using nuclear extract from normal human hepatocytes (hNHeps). A bracket (]) indicates the DNA protein complex. These results were representative of three experiments. IRF, interferon regulatory factor; RXR, retinoid X receptor; MRP2, multidrug resistance protein 2.

We next examined whether these diminished bands represented specific binding to ISRE and attempted to determine which protein was involved in the ISRE–protein complex formation. We used cold competitors and specific antibodies for IRF family members (Fig. 7C). A 50-fold excess of ISRE wild-type cold competitor was sufficient to abolish these bands. However, the retarded bands were not abolished when treated with mutant competitors (Fig. 7C), suggesting that the bands represented specific binding to the ISRE element. All members of the IRF family share homology in the DNA binding domain and bind to a similar DNA motif, the ISRE.24, 25 We performed supershift assays using antibodies of IRF family proteins IRF1, IRF2, IRF3, IRF4, IRF7, IRF8, and IRF9. As shown in Fig. 7C, the retarded bands were abolished specifically only when treated with anti-IRF3 antibody. Other IRF antibodies, Sp1 and RXR, retinoid X receptor (RXR) antibody, failed to shift these bands (Fig. 7C). This suggested that these two bands contain IRF3 protein–DNA complexes. IRF3 thus seemed to bind specifically to the ISRE element of the MRP2 promoter, and IRF3 binding activity to ISRE seemed to be reduced by treatment with IL-1β. We next examined whether nuclear extract from human primary hepatocytes instead of HepG2 could bind to ISRE elements. We observed that the specific ISRE–protein complex bands from human primary hepatocyte were reduced by treatment with IL-1β (Fig. 7D). We also observed that these retarded bands were abolished specifically when pretreated with anti-IRF3 antibody. These results suggested that human primary hepatocytes contain the IRF3 protein–DNA complexes.

Decreased IRF3 Translocation Into the Nucleus on Treatment With IL-1β.

The reduction in IRF3 binding to ISRE could account for the observed IL-1β-mediated downregulation of MRP2 mRNA expression. Reduced binding to ISRE could be caused by a lower expression of IRF3 at the protein level, or by the modification of IRF3 affecting DNA binding or nuclear import. We therefore investigated the expression of IRF3 after IL-1β treatment in the entire cell fraction (Fig. 8A, left), nuclear fraction (Fig. 8A, middle), and cytoplasmic fraction of HepG2 cells (Fig. 8A, right). IRF3 is a phosphoprotein that is constitutively expressed in two forms of approximately 55 kD26 in unstimulated HepG2 cells (Fig. 8A). The upper band of IRF3 could be basal IRF3 phosphorylation corresponding to the N-terminal of IRF3. In the nucleus, IRF3 protein abundance was reduced in this compartment when cells were treated with IL-1β for 8 hours. No reduction in cytoplasmic IRF3 protein and no change in the level of phosphorylation was observed when treated with IL-1β. High mobility group protein-I and glucose-6-phosphate dehydrogenase were used as markers for nuclear and cytoplasmic fraction, respectively. Pretreatment with PD98059 almost completely abolished the IL-1β-induced reduction of nuclear IRF3 protein level, even when cells were treated with IL-1β for 8 hours (Fig. 8B). IL-1β treatment did not change IRF3 protein levels in the whole cell fraction (Fig. 8A,B). To investigate the potential action of IRF3 on MRP2 promoter activity, we used IRF3 expression vector fused Flag tags. The Flag-IRF3 or Flag-only vector was transiently cotransfected into HepG2 cells with a MRP2 promoter luciferase construct for 48 hours. Introduction of Flag-IRF3 from 20 to 200 ng induced an approximately 1.5- to 2.0-fold increase in wild-type MRP2 promoter activity (Fig. 9). By contrast, no induction by Flag-IRF3 was observed when the ISRE mutant reporter was assayed. These results suggested that IRF3 may be involved in the basal transcriptional activity of the MRP2 promoter through the ISRE element.

Figure 8.

Alteration of interferon regulatory factor 3 (IRF3) protein levels in HepG2 in response to interleukin-1β (IL-1β) administration with or without PD98059. 50 μg of whole cell lysate (left panel), 30 μg of nuclear extracts (center), and 50 μg of cytoplasmic extracts (right) were loaded onto each lane. (B) 20 μM PD98059 was administrated for 30 minutes before IL-1β stimulation. Anti-HMG-I antibodies and anti-glucose-6-phosphate dehydrogenase antibodies were used as nuclear and cytoplasmic protein controls. This result was representative of three experiments. HMG-I, high mobility group protein I.

Figure 9.

Alteration of MRP2 promoter-luciferase activity by expression of rIRF3 in HepG2 cells. Cells were transiently transfected with each construct, pRL-TK, and Flag or Flag-IRF3 for 24 hours and assayed for luciferase activity. Luciferase activities were corrected for differences in transfection efficiency by Renilla luciferase activity and were normalized to the activity of the pGL3 basic vector transfected into cells. Each dose of Flag or Flag-IRF3 was transiently transfected. This result was representative of three experiments. *P < .01. IRF3, interferon regulatory factor 3.


Cholestasis and hyperbilirubinemia are common clinical features in various hepatic diseases, including virus-induced or drug-induced hepatitis, alcoholic hepatitis, sepsis, and idiopathic postoperative cholestasis.27, 28 Endotoxins and proinflammatory cytokines are thought to be among the main mediators of inflammation-induced cholestasis.29, 30 Recent studies have reported on the manner by which bile salts or cholesterol homeostasis-regulatory mechanisms are controlled through enterohepatic circulation at the molecular level. Altered expression levels of various transporters, including ABC transporters, have been implicated in cholestasis.31 MRP2 is a typical canalicular ABC transporter that exports many essential organic anions. We previously characterized the 5′-flanking region of the human MRP2 gene and its promoter activity in human hepatic cells.4, 20 Our subsequent study19 demonstrated that MRP2 promoter (p-2635)-driven luciferase activity in hepatic cells was greatly inhibited by IL-1β, and less so by TNFα or IL-6. Consistent with this study, our present study also indicates that MRP2 promoter activity is highly susceptible to the inflammatory cytokine IL-1β. In this study, we demonstrate that treatment of HepG2 cells with IL-1β results in decreased levels of MRP2 mRNA transcription. In addition, we demonstrate that a regulatory protein, IRF3, is responsible for the IL-1β-induced downregulation of MRP2. IRF3 activates the MRP2 promoter through the ISRE under unstimulated conditions. Treatment with IL-1β induces a decrease in both nuclear accumulation of IRF3 and IRF3 binding to ISRE, resulting in downregulation of the MRP2 gene.

Other incidences of IL-1β inducing downregulation of transporters have been reported recently. Denson et al.32 reported that rat ntcp and mrp2 genes are regulated by retinoic acid receptor:RXR, and that IL-1β reduces nuclear retinoic acid receptor:RXR. The retinoic acid receptor:RXR site is located in −210/−180 bp upstream from the transcription initiation site in the human MRP2 promoter.

Our promoter deletion analysis and mutation of the promoter show that the IL-1β-induced inhibition of MRP2 promoter activity is the result of the ISRE region downstream of retinoic acid receptor:RXR. It has also been reported that IL-1β-induced downregulation of the rat ntcp gene occurs via the JNK pathway,33 whereas in our study, the downregulation of human MRP2 by IL-1β occurs via the ERK pathway. Kast et al.34 also reported that rat mrp2 expression is regulated by the pregnane X receptor and constitutive androgen receptor, and that rat mrp2 promoter activity is activated by nuclear receptor pregnane X receptor, constitutive androgen receptor, and their agonist. These results suggest that MRP2 expression in the liver is regulated by environmental stimuli, numerous compounds through various nuclear receptors, and IRF3. The 5′-flanking region of the MRP2 promoter contains C/EBPβ, HNF1, and USF elements, as well as an ISRE (Fig. 5).20 Mutant constructs of ISRE and HNF show decreasing basal promoter activities when compared with the wild-type MRP2 promoter, suggesting that these elements play a key role in the basal promoter activity. By contrast, there is an almost 3.0-fold increase in the promoter activity by a C/EBPβ mutation in the context of the MRP2 promoter over wild type. Thus, C/EBPβ may regulate MRP2 promoter activity negatively. We observed that IL-1β-induced suppression of MRP2 mRNA is associated with the ISRE region of the MRP2 promoter (Fig. 5). Among the various factors interacting with ISRE, IRF3 was found to bind specifically to the ISRE on the MRP2 promoter (Fig. 7). EMSAs show two bands containing IRF3–protein complexes. These complexes may be the result of different proteins interacting with IRF3 or conformational changes within the protein. Expression of the exogenous IRF3 gene increases MRP2 promoter activity through the ISRE and when cells are treated with IL-1β levels of IRF3 in the nucleus are diminished (Fig. 8). IRF3 thus seems to be a positive transcription factor of the MRP2 promoter in hepatic cells, and decreased levels of IRF3 in the nucleus may be responsible for IL-1β-induced downregulation of the MRP2 promoter activity. However, we could not ruled out other transcription factor in IL-1β-mediated suppression of MRP2.

Infection of fibroblasts with human cytomegalovirus also causes nuclear translocation of IRF3 and cooperative DNA binding with the transcriptional coactivator CBP/p300.35 On infection of a cell with a virus, the C-terminal domain of cytoplasmic IRF3 is phosphorylated and translocated to the nucleus of hematopoietic cells, where it activates several virus-induced genes.26 Our work shows that IRF3 is expressed continuously in both the nucleus and cytoplasm of hepatic cells (Fig. 8). Uchiumi et al. (unpublished data). have demonstrated by immunohistochemical study that IRF3 is found to be localized mainly to the cytoplasm, with 10% of whole cell IRF3 in the nucleus. We observe a marked decrease in IRF3 protein levels in the nucleus accompanied by reduced formation of the IRF3–DNA complex when cells are treated with IL-1β for 6 hours or longer.

Coadministration of an ERK inhibitor (PD98059) relieves the suppression of MRP2 mRNA expression and MRP2 promoter activation by IL-1β and reverses the decreased binding of IRF3 to ISRE seen in the presence of IL-1β. IL-1β seems to decrease nuclear levels of a positive regulatory factor, IRF3, resulting in suppression of the MRP2 promoter through the loss of interaction with ISRE. This process of IL-1β-induced suppression of the MRP2 gene could be linked closely with ERK1/2 activation signaling.

In conclusion, among the many interferon regulatory factors, we determined that IRF3 has a specific interaction with ISRE on the MRP2 promoter. The expression of exogenous IRF3 in hepatic cells activates reporter gene expression via the p-491MRP2 promoter construct containing wild-type ISRE, but not by p-491MRP2 (ISREmt) containing mutant ISRE. The expression levels of IRF3 in the nucleus decrease after cells are treated with IL-1β for 8 hours, concomitant with a reduction in the binding of IRF3 to ISRE at the MRP2 promoter.