Pregnane X Receptor PXR Activates the GADD45β Gene, Eliciting the p38 MAPK Signal and Cell Migration

Abstract

Pregnane X receptor (PXR) was originally characterized as a transcription factor that induces hepatic drug metabolism by activating cytochrome P450 genes. Here we have now demonstrated a novel function of PXR, that of eliciting p38 mitogen-activated protein kinase (MAPK) phosphorylation for cell migration. Upon xenobiotic activation of ectopic human PXR, human hepatocellular carcinoma HepG2 cells were found to exhibit increased phosphorylation of p38 MAPK and to subsequently change morphology and migrate. p38 MAPK was responsible for the regulation of these morphological changes and cell migration because the p38 MAPK inhibitor SB239063 repressed both. Prior to this phosphorylation, PXR directly activated the early response GADD45β gene by binding to a distal direct repeat 4 site of the GADD45β promoter. Ectopic expression of GADD45β increased p38 MAPK phosphorylation, whereas siRNA knockdown of GADD45β decreased the PXR-induced p38 MAPK phosphorylation, confirming that GADD45β can regulate PXR-induced p38 MAPK phosphorylation in HepG2 cells. These results indicate that PXR activates the GADD45β gene, increasing p38 MAPK phosphorylation, and leading HepG2 cells to change morphology and migrate. The GADD45β gene is a direct target for PXR, eliciting cell signals to regulate various cellular functions.

Introduction

Pregnane X receptor (PXR²; NR1I2), an orphan member of the nuclear steroid/thyroid receptor superfamily, was originally characterized as the xenobiotic-activated transcription factor. The role first established for PXR was to regulate the xenobiotic response activation of genes that encode xenobiotic-metabolizing enzymes and transporters, thus increasing metabolism and excretion of xenobiotics (1). Subsequently, PXR was also found to regulate hepatic energy metabolism, wherein PXR interacts with insulin/glucagon-responsive factors such as FoxO1, FoxA2, PGC1, and CREB to repress hepatic genes such as G6Pase, Pepck1, Cpt1a, and Hmgcs, thereby attenuating glucogenogenesis, fatty acid oxidation, and ketogenesis (2–5). In addition, recent studies have shown that PXR also regulates vitamin D metabolism. Through the regulation of these metabolisms, PXR can be playing critical roles in the development of various types of metabolic diseases such as hepatic hypertrophy, acetaminophen and bilirubin toxicities, steatosis, cholestasis, diabetes, and osteomalacia (6–10). Although these metabolic roles and the clinical implications of PXR have now been established, a nonmetabolic role of PXR in the regulation of cellular signals has now begun to emerge. For example, NF-κβ-mediated inflammatory signals and bowel inflammation was up-regulated in Pxr^−/− mouse (3). PXR represses drug-induced apoptosis in various cell systems, human and rat primary hepatocytes and human colon cancer HTC cells (11, 12). The key questions, however, remain unanswered at the present time: what are the direct targets of PXR that initiate signals and what is the molecular mechanism by which PXR regulates these targets?

GADD45β, is an immediate-early response gene induced by various physiological and environmental stressors, including cytokines and genotoxic stresses (13). Through protein-protein interactions with numerous signal molecules, GADD45β can regulate cellular signals for cell cycle, DNA repair, and apoptosis, depending on the types of stimuli and cells that are stimulated (14–17). For example, GADD45β activates the p38 mitogen-activated protein kinase (MAPK) signal pathway via direct interaction with MTK1/MEKK4, a MAPK kinase kinase (15, 17), repressing angiogenesis in the human pancreatic carcinoma Panc-1 cells (18).

We utilized human hepatocellular carcinoma HepG2 cells stably expressing human PXR (called ShP51) and observed that activation of PXR by an antibiotic rifampicin (RIF) stimulated phosphorylation of p38 MAPK. Subsequent microarray analysis identified the GADD45β gene as the gene induced prior to activation of p38 MAPK signal pathway immediately after RIF treatment in HepG2 cells. Therefore, we investigated the molecular mechanism in which PXR elicits a p38 MAPK signal by directly activating the GADD45β gene. PXR bound directly to the newly identified response element within the GADD45β promoter and activated it in cell-based transcription assays. An overexpression of GADD45β resulted in the increase phosphorylation of p38 MAPK in HepG2 cells, whereas siRNA knockdown of GADD45β decreased PXR-dependent phosphorylation of p38 MAPK in ShP51 cells. ShP51 cells were also found to change morphology and migrate after RIF treatment. Chemical inhibition of p38 MAPK repressed these morphological changes and migration. Our present study has clearly demonstrated a novel PXR function and has provided us with the basis to investigate the molecular mechanism by which PXR regulates cell signals and fates such as morphology and migration. Because humans are constantly exposed to numerous therapeutics and xenobiotics, a PXR-elicited cell signal may become a critical factor in understanding human susceptibility to the toxicity and carcinogenicity caused by chemical exposures.

EXPERIMENTAL PROCEDURES

Materials

Restriction endonucleases and DNA-modifying enzymes were from New England Biolabs (Beverly, MA). Mouse monoclonal antibody to human PXR was from Perseus Proteomics Inc. (Tokyo, Japan). Mouse monoclonal antibody to V5 was from Invitrogen. Antibodies to phospho-p38 MAPK (Thr¹⁸⁰/Tyr¹⁸²), p38 MAPK, phospho-MKK3/6 (Ser¹⁸⁹/Ser²⁰⁷), MKK6, phospho-MK2 (Thr³³⁴), MK2, phospho-JNK (Thr⁸³/Tyr¹⁸⁵), JNK, phospho-ERK1/2 (Thr²⁰²/Tyr²⁰⁴), ERK1/2, phospho-AKT (Ser⁴⁷³), AKT, phospho-AMPK (Thr¹⁷²), AMPK, and phospho-c-Jun (Ser⁶³) were from Cell Signaling Technology (Beverly, MA). Normal mouse IgG, anti-GADD45β, c-Jun, and β-actin antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA). [³²P]dATP was from GE Healthcare.

Vectors

pCR3/hPXR, pcDNA3.1/hPXR, pCMX/hRXR, pCR3/hPGC1α, adeno-hPXR, and adeno-β-galactosidase were described previously (5). A full-length cDNA of human GADD45β, that was cloned from HepG2 with the use of primer pair 5′-GGTACCATGACGCTGGAAGAGCTCGTGGCG-3′ and 5′-CTCAGCGTTCCTGAAGAGAGATGTAGGGG-3′, was inserted into pcDNA3.1/V5-His-TOPO (Invitrogen) to produce pcDNA3.1/hGADD45β. pCR3/hPXR_R98C and pcDNA3.1/hPXR_R98C were generated from pCR3/hPXR and pcDNA3.1/hPXR, respectively, by site-directed mutagenesis using the QuikChange mutagenesis kit (Stratagene, La Jolla, CA) and a proper pair of mutagenic oligonucleotides. Human PXR_R98C cDNA, digested from pCR3/hPXR_R98C, was inserted into pAdtrackCMV vector (American Type Culture Collection, Manassas, VA) to produce adeno-hPXR_R98C. The −1565/+47 region of the human GADD45β promoter was amplified from human genomic DNA (Promega, Madison, WI) using LA Taq DNA polymerase (TaKaRa, Ohtsu, Japan) and the following primers: 5′-GAGATCTGTATGTTGCATGCGTAAAACATTGCAT-3′ and 5′-CAAGCTTGCGAGGATAATCCAGGAAGTTGCGG-3′. Amplified DNA was digested with BglII and HindIII and was inserted into pGL3-Basic (Promega) to yield pGL3/hG45β. To generate pGL3/hG45βcore containing the −66/+47 region of GADD45β promoter, a XhoI site was inserted into pGL3/hG45β by site-directed mutagenesis, following digestion with XhoI and self-ligation. A double-stranded oligonucleotide 5′-CAGGCAGATCATTTGAGGTCAGGAGAGGCAGATCATTTGAGGTCAGGAGAGGCAGATCATTTGAGGTCAGGAGC-3′ was inserted into the XhoI site of pGL3/hG45βcore to generate pGL3/3×DR4hG45βcore.

Cell Culture, Drug Treatment, Transfection, and Infection

Human hepatocellular carcinoma HepG2 and Huh7 cells were maintained in MEM supplemented with 10% FBS, 2 mm l-glutamine, penicillin (100 units/ml), and streptomycin (100 μg/ml) in an atmosphere of 5% CO₂ at 37 °C. For qRT-PCR and Western blotting, respectively, total RNAs and whole cell lysates were prepared from HepG2 cells that were treated with 10 μm RIF (Sigma-Aldrich) or 3 μm SR12813 (Sigma-Aldrich) in FBS-free MEM for a given time. For luciferase reporter assays, Huh7 cells were co-transfected with human GADD45β promoter-firefly luciferase, pRL-CMV for Renilla luciferase control (Promega), and pCR3/hPXR or pCR3/hPXR_R98C, using FuGENE 6 (Roche Applied Science). The final amounts of transfected DNAs were adjusted by adding pcDNA3.1-V5-His as empty vector control. Twenty four hours after transfection, these cells were subsequently treated with a given drug in FBS-free MEM for an additional 24 h. Luciferase reporter activities were measured as described previously (2). For ectopic expression of GADD45β, trypsinized HepG2 cells were reverse-transfected with increasing amount of pcDNA3.1/hGADD45β, using FuGENE 6. The final amounts of transfected DNAs were adjusted by adding pcDNA3.1-V5-His. After 30 h, whole cell lysates were prepared. For adenoviral infection, HepG2 cells were cultured in MEM containing adeno-β-galactosidase, adeno-hPXR, or adeno-hPXR_R98C at 10 of multiple of infection. After 30 h, these cells were treated with 10 μm RIF or 3 μm SR12813 in FBS-free MEM for a given time. Then, total RNAs and whole cell lysates were prepared. For siRNA knockdown, trypsinized HepG2 cells were reverse-transfected with 40 μm ON-TARGETplus SMART pool GADD45β (catalogue number L-003894-00) or ON-TARGETplus siCONTROL nontargeting pool (catalogue number D-001810-10) from Dharmacon Research (Lafayette, CO) in MEM for 48 h, using Lipofectamine 2000 (Invitrogen). Then, these cells were treated with dimethyl sulfoxide (DMSO) or RIF in FBS-free MEM for 1 h, from which total RNAs and whole cell lysates were prepared for qRT-PCR and Western blotting, respectively. For knockdown of p38 MAPK, cells were reverse-transfected with ON-TARGETplus SMART pool p38 MAPK (catalogue number L-003512-00) for 36 h and were subsequently treated with RIF for a given time.

ShP51 Cells That Stably Express Human PXR

HepG2 cells were transfected with pCR3/hPXR by FuGENE 6 and were selected in MEM containing G418 (Invitrogen) at a concentration of 800 μg/ml. Drug-resistant colonies were further selected and verified by Western blotting of PXR and qRT-PCR of CYP3A4 to establish ShP51 cells.

Western Blotting

Cells were lysed and denatured in a fixed volume of NuPAGE LDS sample buffer (Invitrogen), from which a fixed volume was separated on a 8.5%, a 10%, or a 10% SDS-polyacrylamide gel and were transferred onto PVDF membrane. This membrane was blocked with 5% milk in TBS-T for 1 h at room temperature and then incubated with a given primary antibody in TBS-T containing 5% BSA for additional 16 h at 4 °C, prior to the incubation with secondary antibody in TBS-T with 5% milk for 2 h at room temperature. Immunoreactive bands were visualized using ECL plus Western blotting detection reagents (GE Healthcare).

Real-time PCR

Total RNAs were extracted using TRIzol reagent (Invitrogen) to synthesize cDNA using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). qRT-PCR was performed with an ABI prism 7700 sequence detection system (Applied Biosystems). Assays-on-demand probes (Applied Biosystems) were used for PCR with the TaqMAN PCR Master Mix (Applied Biosystems), Hs00430021_m1 for the human CYP3A4 gene. The following PCR primers were used with the SYBR Green Master Mix (Applied Biosystems): hGADD45β-RT-S, 5′-CGGTGGAGGAGCTTTTGGT-3′ and hGADD45β-RT-AS, 5′-GCTGTCTGGGTCCACATTCA-3′ for the human GADD45β gene. The TaqMAN human β-actin control regents kit (Applied Biosystems) was used as the internal control.

Gel Shift

cDNAs for human PXR in pCR3/hPXR, human PXR_R98C in pCR3/h PXR_R98C and human RXR in pCMX/hRXR were in vitro translated by using a TNT-coupled reticulocyte lysate system (Promega). The double-stranded probes used for hGADD45β-DR4 and CYP3A4-ER6 were 5′-GATCAGGCAGATCATTTGAGGTCAGGAG-3′ and 5′-GATCATATGAACTCAAAGGAGGTCAGTG-3′, respectively, which were labeled by using [³²P]dATP and DNA polymerase Klenow fragment. The double-stranded probe 5′-GATCATAGGAACGCAAAGGCGGTCCGTG-3′ (CYP3A4-ER6mt) was used for competition assays. Using these proteins and probes, gel shift assays were performed as described previously (19). For competition and supershift assays, unlabeled probe, normal mouse IgG, and anti-human PXR antibody were preincubated for 15 min before adding the radioactive probes to start reactions.

ChIP Assay

ChIP assay was performed using a ChIP assay kit (Millipore). Briefly, trypsinized HepG2 cells were reverse-transfected with pcDNA3.1-V5-His, pcDNA3.1/hPXR, or pcDNA3.1/hPXR_R98C, using FuGENE 6. After 48 h, these cells were treated with RIF in FBS-free MEM for 1 h, cross-linked by adding formaldehyde (final concentration of 1% (v/v)) in medium, and incubated for 10 min at 37 °C. Pellets of these cross-linked cells were sonicated to shear DNA in the SDS lysis buffer on wet ice. After being precleared by shaking with protein A agarose, these lysates were incubated with anti-V5 antibody or normal mouse IgG at 4 °C overnight. The immunoprecipitated DNA was purified by following the manufacturer's instructions combined with QIAquick PCR purification kit (Qiagen, Valencia, CA) and was used as a template for semiquantitative PCR with LA Taq polymerase (TaKaRa). The following primers were used for PCR: for hGADD45β-DR4, 5′-CCAAATACGGTGACTCACGCCTGTAATTC-3′ and 5′-GTAGAGACAGGGTTTTGCCATGTTG-3′; for hGADD45β-nega, 5′-GCACAATCTCGGCTCACTGCAACC-3′ and 5′-GATCGAGACCATCCTGGCCAACATG-3′; for CYP3A4-XREM, 5′-ACTCATGTCCCAATTAAAGGTC-3′ and 5′-TGTTCTTGTCAGAAGTTCAGC-3′ (20).

Cell Migration

For migration assay, cells were seeded onto the upper surface of dividing membrane (8 μm pore size) of a Transwell Boyden chamber (Corning Inc., Corning, NY). After 24 h, the upper chambers were loaded with FBS-free MEM with or without 10 μm SB239063 and FBS-free MEM with or without 10 μm RIF, and 10 μm SB239063 (Sigma-Aldrich) was added to the lower well. After incubation for 48 h, the membranes were fixed in PBS containing 5% (v/v) formaldehyde and stained with a 0.1% (w/v) crystal violet solution. The number of cells migrated to the reverse surface of the membrane was counted in three randomly selected fields under light microscopy.

Cell Staining

Forty eight hours after seeding, cells were incubated in FBS-free MEM with drugs for another 48 h. These cells were fixed in PBS containing 5% (v/v) formaldehyde, followed by staining with a 0.1% (w/v) crystal violet solution. To see actin organization, these drug-treated cells were fixed in PBS containing 5% (v/v) formaldehyde, incubated in PBS containing 0.1% (v/v) Triton X-100 and washed with PBS, followed by staining with Alexa Fluor 568 phalloidin (Invitrogen) and by observing the cells under a confocal microscopy. Additional procedures may be found in supplemental Methods.

RESULTS

PXR Elicits a MKK3/6-p38 MAPK Signal

Twenty independent HepG2 cells stably expressing human PXR were cloned via neomycin resistance. As shown in one of them, named ShP51 cells, all of these stable cells were confirmed to express PXR protein and to activate its typical target CYP3A4 gene after treatments with two classic activators RIF and SR12813 for 24 h (Fig. 1A). Western blot analysis of whole cell lysates prepared from these same parental HepG2 and ShP51 cells revealed that phosphorylation of p38 MAPK was greatly increased after drug treatments (Fig. 1B). Given this finding, time-dependent increases of phosphorylation of p38 MAPK and its upstream and downstream kinases were examined; MAPK kinase (MKK) 3/6 and MAPK-activated protein kinase 2 (MK2), respectively (Fig. 1C). In ShP51 cells after being treated with RIF, phosphorylation of p38 MAPK was dramatically increased as early as 30 min and reached an apparent peak at 60 min. This high level of this p38 MAPK phosphorylation remained 240 min after RIF treatment. Increase of phosphorylation of MKK3/6 was found to be as fast as that of p38 MAPK after RIF treatment, whereas MK2 exhibited a delayed increase of phosphorylation compared with those of MKK3/6 and p38 MAPK. Phosphorylation of MK2 was first increased 60 min after RIF treatment and thereafter remained increased in the ShP51 cells. Any of these phosphorylations were not observed in RIF-treated parental HepG2 cells. These results indicate that PXR activated MKK3/6-p38 MAPK-MK2 signal in response to drug treatment. Subsequently, phosphorylation levels of JNK1/2, ERK1/2, AKT, and AMPK were investigated to examine specificity of PXR-activated phosphorylation in ShP51 cells (supplemental Fig. S1). Phosphorylation of JNK1/2 was increased, but this increase was weak until 120 min after RIF treatment. Those of ERK1/2, AKT, and AMPK were not increased in any time of RIF treatment. Thus, the PXR-induced increase of p38 MAPK was a quick and relatively specific response to RIF treatment in ShP51 cells.

FIGURE 1.

PXR elicits p38 MAPK phosphorylation. A, establishment of ShP51 cells. Ectopic expression of PXR was verified by Western blotting of whole cell lysates using anti-human PXR antibody and by qRT-PCR of the CYP3A4 mRNA in the ShP51 cells treated with the two distinct PXR activators, RIF and SR12813. The CYP3A4 mRNA levels were expressed by taking those in the DMSO-treated cells as 1. Columns represent the mean ± S.D. (error bars). B, phosphorylation of p38 MAPK by PXR activators. After a 4-h treatment with either RIF or SR12813, whole cell lysates were subjected to Western blotting with antibodies for p38 MAPK and actin. C, time response of p38 MAPK phosphorylation. At each time point after RIF treatment, whole cell lysates were subjected to Western blotting using given antibodies: p38 MAPK, MKK3/6, MK2, and actin.

PXR Activates the GADD45β Gene

Because PXR is a ligand-dependent transcription factor, its primary targets are genes. We asked whether PXR elicited p38 MAPK phosphorylation via activating a gene. Naturally occurring mutation of arginine 98 to cysteine abolished DNA-binding ability of PXR (21). Utilizing this mutation, we examined whether or not PXR required its DNA-binding ability to elicit phosphorylation of p38 MAPK. For this purpose, wild-type PXR and its PXR_R98C mutant were ectopically expressed in HepG2 cells. Consisting with the lack of DNA binding, adenovirus-based expressing PXR_R98C did not activate the expression of CYP3A4 gene in HepG2 cells following RIF treatment (Fig. 2A). As can be seen in Fig. 2B, PXR_R98C was not able to increase the phosphorylation levels of both MKK3/6 and p38 MAPK after RIF treatment (Fig. 2B). These results suggested that PXR increased phosphorylation of p38 MAPK by activating a gene in RIF-treated HepG2 cells.

FIGURE 2.

PXR_R98C mutant that lacks DNA binding fails to elicit a MKK3/6-p38 MAPK signal in HepG2 cells. A, PXR_R98C mutant lacking CYP3A4 induction. HepG2 cells were infected with adeno-β-galactosidase, adeno-hPXR, or adeno-hPXR_R98C for 30 h; after that RIF was added for another 24 h. Subsequently, total RNAs were prepared and subjected to qRT-PCR. The CYP3A4 mRNA levels were expressed by taking those in the DMSO-treated cells infected with adeno-β-galactosidase as 1. Columns represent the mean ± S.D. (error bars). B, no elicitation of MKK3/6-p38 MAPK signal by PXR_R98C mutant. After adenoviral infection, HepG2 cells were treated with RIF for 2 h. Subsequently, whole cell lysates were prepared and subjected to Western blotting using given antibodies: MKK3/6, p38 MAPK, PXR, and actin.

To identify the PXR-activated gene required for the phosphorylation of p38 MAPK, we initially performed microarray analysis (supplemental Table S1), from which the immediate-early response GADD45β gene was selected for further investigation. This selection was made based on the fact that GADD45β is known to regulate phosphorylation of p38 MAPK (15, 17, 18) and the hypothesis that the gene needs to quickly respond to PXR activation to stimulate p38 MAPK phosphorylation within 30 min following RIF treatment. Increase of GADD45β mRNA was, in fact, detected as early as 15 min and peaked 60 min following RIF treatment in ShP51 cells but not in parent HepG2 cells (Fig. 3A). The GADD45β mRNA remained at this level for 24 h following RIF treatment (data not shown). This quick response of the GADD45β gene became more evident when it was compared with a slower increase of the CYP3A4 mRNA, the classic PXR-targeted gene: the most significant increase in CYP3A4 mRNA occurred between 60 and 120 min after RIF treatment. However, the RIF-induced increase of GADD45β mRNA was not observed in HepG2 cells infected with adenovirus expressing PXR_R98C mutant, substantiating the notion that PXR binds directly to the GADD45β gene and activates its transcription (Fig. 3B). Hepatocellular carcinoma Huh7 cells were utilized to confirm that the observed PXR-mediated increase of GADD45β mRNA and of phosphorylation of p38 MAPK was not specific to HepG2 cells (supplemental Fig. S2). These results suggested that GADD45β was the direct target of PXR to elicit p38 MAPK signal.

FIGURE 3.

Drug activation of PXR induces the GADD45β gene. A, increase in GADD45β mRNA levels after RIF treatment. Cells were harvested at each time point after RIF treatment, from which total RNAs were prepared and subjected to qRT-PCR. These GADD45β and CYP3A4 mRNA levels were expressed by taking those in the DMSO-treated cells as 1. Columns represent the mean ± S.D. (error bars). B, PXR_R98C mutant. HepG2 cells were infected with adeno-β-galactosidase, adeno-hPXR, or adeno-hPXR_R98C for 30 h. After that RIF was added for another 2 h. Subsequently, total RNAs were prepared and subjected to qRT-PCR. These mRNA levels were expressed by taking these in DMSO-treated HepG2 cells infected with adeno-β-galactosidase as 1. Columns represent the mean ± S.D.

Direct activation by PXR of the GADD45β gene was further investigated by performing gel shift, luciferase reporter, and ChIP assays. Motif analysis revealed a direct repeat 4 (DR4) sequence (⁻⁴³⁸⁶AGATCATTTGAGGTCA⁻⁴³⁷¹) within a 10-kb GADD45β promoter as a putative site for PXR binding. Gel shift assays showed a specific binding of PXR·RXR complex to the DR4 sequence which was confirmed by supershift using anti-human PXR antibody (Fig. 4A). A triple repeat of the DR4 sequence was placed in front of the −66/+47 bp of the hG45βcore-Luc reporter plasmid which contains the GADD45β proximal promoter to produce 3×DR4-hG45βcore-Luc. These reporter plasmids were co-transfected with either PXR or PXR_R98C expression plasmid into Huh7 cells. PXR activated the 3×DR4-hG45βcore-Luc, but not the hG45βcore-Luc in RIF-treated Huh7 cells (Fig. 4B). PXR_R98C did not activate either 3×DR4-hG45βcore-Luc or hG45βcore-Luc. Finally, ChIP assays confirmed the binding of PXR to the DR4 sequence of the GADD45β promoter, in which V5-tagged PXR and PXR_R98C were ectopically expressed in RIF-treated HepG2 cells. PXR, but not PXR_R98C, clearly bound to the DR4 of the GADD45β promoter, similar to those observed in the CYP3A4 promoter (Fig. 4C). In contrast, PXR did not bind to a region lacking the DR4. These results indicate that PXR can directly activate the GADD45β gene.

FIGURE 4.

PXR directly activates the GADD45β gene. A, gel shift and supershift assays. A ³²P-labeled DR4 sequence (⁻⁴³⁸⁶AGATCATTTGAGGTCA⁻⁴³⁷¹ from the GADD45β promoter) was incubated with the in vitro translated PXR or PXR_R98C in the presence or absence of RXR. The resulted DNA·PXR complex was analyzed on a polyacrylamide gel. Anti-human PXR antibody, mouse normal IgG, unlabeled oligonucleotide, and its mutant were used to verify the specific formation of DNA·PXR complex. The ³²P-labeled ER6 oligonucleotide from the CYP3A4 promoter was utilized as a position control for the assays. B, luciferase reporter assays. pCR3/hPXR or pCR3/hPXR_R98C was co-transfected with pGL3-Basic, pGL3/hG45βcore, pGL3/3×DR4hG45βcore, and pRL-CMV into Huh7 cells. Twenty four hours after RIF treatment, these cells were subjected to luciferase reporter assays. Independently, XREM-3A4-Luc was utilized as the positive controls for trans-activation activity of PXR. Relative -fold of activity was calculated by taking the activity of cells co-transfected by pcDNA3.1-V5-His with pGL3-Basic in the presence of DMSO as 1. Columns represent the mean ± S.D. (error bars). C, ChIP assays. V5-tagged PXR or PXR_R98C was ectopically expressed in HepG2 cells, which were treated with RIF for 1 h and subjected to ChIP assays using anti-V5 antibody. From the immunoprecipitated DNA fragments, both the regions containing and lacking the DR4 sequence were amplified by PCR for the GADD45β promoter; GADD45β-DR4, and GADD45β-nega. The region containing the ER6 sequence was also amplified for the CYP3A4 promoter as positive control, CYP3A4-XREM. An equal expression of PXR and PXR_R98C was verified by Western blotting of the whole cell lysates with anti-V5 antibody.

GADD45β Mediates PXR-elicited p38 MAPK Phosphorylation

To examine whether GADD45β was responsible for RIF-induced increase of p38 MAPK phosphorylation, GADD45β was either overexpressed in HepG2 cells or knocked down in ShP51 cells. Ectopic GADD45β was expressed in HepG2 cells by transfecting the GADD45β expression plasmid in a dose-dependent manner. Western blotting showed a barely detectable expression of GADD45β at 0.1 and 0.3 μg of transfected plasmid and a high expression at 0.9 μg (Fig. 5A). Elevated phosphorylation of MKK3/6 and p38 MAPK was already detected at 0.1 μg, which was further increased at 0.3 μg. These phosphorylations were saturated at 0.3 μg and were no longer increased at 0.9 μg. This phosphorylation of MKK3/6 and p38 MAPK, thus, appeared to be very sensitive to GADD45β. Similar results were also obtained in Huh7 cells (data not shown). Next, we utilized siRNA to knock down endogenous GADD45β in ShP51 cells to ascertain whether GADD45β can determine a PXR-dependent p38 MAPK phosphorylation. Transfection of GADD45β siRNA, but not control siRNA, repressed PXR-elicited phosphorylation of MKK3/6 and p38 MAPK in ShP51 cells following RIF treatment (Fig. 5B). In those cells, GADD45β siRNA specifically decreased both basal and induced levels of GADD45β mRNA, but not those of CYP3A4 mRNA. GADD45β, thus, mediated PXR-dependent phosphorylation of MKK3/6 and p38 MAPK, suggesting that PXR activated the GADD45β gene and the induced GADD45β, in turn, stimulated p38 MAPK signal.

FIGURE 5.

GADD45β regulates PXR-elicited p38 MAPK phosphorylation. A, overexpression. HepG2 cells were reverse-transfected with increasing amounts of pcDNA3.1/hGADD45β for 30 h, from which whole cell lysates were prepared for subsequent Western blotting using given antibodies: GADD45β, MKK3/6, p38 MAPK, and actin. B, siRNA knockdown. ShP51 cells were reverse-transfected with control or GADD45β siRNAs for 48 h and were treated with RIF for an additional 1 h. Whole cell lysates were prepared for subsequent Western blotting using given antibodies: MKK3/6, p38 MAPK, and actin. Total RNAs were prepared and subjected to qRT-PCR. These mRNA levels were expressed by taking these DMSO-treated ShP51 cells transfected with control siRNA as 1. Columns represent the mean ± S.D. (error bars).

PXR-GADD45β-p38 MAPK Signals Migrate HepG2 Cells

PXR, upon activation by RIF, directly activates the GADD45β gene to elicit p38 MAPK signal in HepG2 cells. Activation of p38 MAPK could have various consequences for cell response, including cell migration and apoptosis. We first observed that ShP51 cells, but not parental HepG2 cells, underwent a striking morphological change wherein cells scattered and flattened upon treatment with RIF or SR12813 (Fig. 6A). The additional 14 clones of HepG2 cell lines stably expressing PXR also changed their morphology in the same manner as ShP51 cells (data not shown). Moreover, adenovirus-based expressing PXR, but not PXR_R98C, caused morphological changes after RIF treatment similar to those observed in ShP51 cells (supplemental Fig. S3). Thus, these results clearly indicated that the observed morphological changes occurred as a consequence of PXR activation by drugs. A subsequent immunohistochemical study showed that these morphological changes were accompanied with the reorganization of actin filaments often observed during cell migration (Fig. 6B): formations of stress fiber, lamellipodia, and filopodia (22, 23). Therefore, transwell migration assays were performed to examine whether or not ShP51 cells would migrate after treatments with RIF and SR12813. Cells that migrated onto the other side of the transwell were visualized, and their numbers were counted in Fig. 6C. ShP51 cells, but not parental HepG2 cells, were found to exhibit a 5-fold increase in migration following treatment with RIF and SR12813. Thus, drug activation of PXR resulted in morphological changes, leading ShP51 cells to migrate. In addition, cell growth was not affected throughout the duration of morphological changes and migration after treatments with RIF and SR12813 (data not shown).

FIGURE 6.

PXR alters morphology, migrating HepG2 cells. A, morphological changes. Cells were treated with DMSO, RIF, or SR12813 for 48 h, fixed, and stained. Scale bar, 100 μm. B, actin reorganization. After the same drug treatments, cells were stained with Alexa Fluor 568 phalloidin. Scale bar, 50 μm. C, migration. Cells were grown on the membrane of a transwell Boyden chamber in the presence of DMSO, RIF, or SR12813 for 48 h. The migrated cells were fixed, stained, and counted. Scale bar, 200 μm. Three independent experiments were performed to average migrations. Columns represent the mean ± S.D. (error bars).

The specific p38 MAPK inhibitor SB239063 was employed to link PXR-mediated morphological changes and cell migration with p38 MAPK in HepG2 cells. Under the assay condition, phosphorylation of an immediate p38 MAPK downstream kinase MK2 was completely inhibited by concentrations as low as 10 μm SB239063 in RIF-treated ShP51 cells (Fig. 7A). Accordingly, ShP51 and parental HepG2 cells were co-treated with RIF and 10 μm SB239063 and were subjected to morphological and transwell migration analyses to examine whether or not p38 MAPK, in fact, regulated PXR-induced morphological changes and migration. No morphological changes were observed in ShP51 cells co-treated with SB239063 (Fig. 7B). A 5-fold increase by RIF treatment of migration of ShP51 cells was completely repressed by co-treatment with SB239063 (Fig. 7C). SB239063 treatment also increased c-Jun phosphorylation in both parental HepG2 and ShP51 cells (Fig. 7A).

FIGURE 7.

p38 MAPK regulates PXR-induced morphology and migration. A, specific inhibition of p38 MAPK activity by SB239063. After being pretreated with increasing amounts of SB239063 for 2 h, cells were treated with RIF for additional 4 h, from which cell lysates were prepared for subsequent Western blotting using these given antibodies: MK2, c-Jun, and actin. B, morphological changes. Cells, which were pretreated with DMSO or SB239063 for 2 h, were co-treated with RIF for additional 48 h prior to fixation and staining. Scale bar, 100 μm. C, migration. Cells were grown on the dividing membrane of a transwell Boyden chamber and were treated with RIF in the presence or absence of SB239063 for 48 h. The migrated cells were visualized and counted. These values are averaged from three independent experiments. Columns represent the mean ± S.D. (error bars).

To confirm the role of the p38 MAPK in RIF-induced morphological changes, ShP51 cells were transfected with siRNA to knock down endogenous p38 MAPK. This knockdown resulted in significant repression of RIF-induced morphological changes (supplemental Fig. S4). Taken together, p38 MAPK was demonstrated to be responsible for the PXR-induced morphological changes and cell migration.

DISCUSSION

Upon drug activation, PXR has now been shown to directly activate the GADD45β gene by binding to a DR4 sequence within its promoter, eliciting a MKK3/6-p38 MAPK signal and migrating HepG2 cells. The PXR regulation of cell migration via the GADD45β gene emphasizes the possibility that PXR may play diverse roles within the realms of cell regulation, because GADD45β is known to regulate various types of cellular functions from apoptosis to cell cycle and DNA repair through its interactions with various signal factors such as the cdc2/cyclin B1 complex and PCNA (13).

GADD45β is rapidly induced by genotoxic/oxidative stresses and cytokines and acts as a stress sensor to coordinate cellular response signals such as cell survival and apoptosis. TGFβ is one such stress-mediated factor that activates the GADD45β gene to promote apoptosis via p38 MAPK signal: the TGFβ-induced apoptosis and p38 MAPK phosphorylation were diminished in the primary hepatocytes prepared from Gadd45β KO mice (24). In our present study with HepG2 cells, induction of GADD45β by PXR had no affect on cell growth under the experimental conditions used; the known TGFβ-induced apoptosis-resistant nature of HepG2 cells could be a factor in this phenomenon (25). Instead, an up-regulated GADD45β-MKK3/6-p38 MAPK signal resulted in epithelial-mesenchymal transition-like morphological changes and migration of HepG2 cells. This finding, however, is consistent with the previous observations that TGFβ causes HepG2 cells to invade (26) and that p38 MAPK mediates TGFβ-induced migration of mouse mammary epithelial NMuMG cells (27). Furthermore, it is recently reported that activation of p38 MAPK signal pathway leads hepatocellular carcinoma to migrate (28). MK2, a well known downstream kinase of p38 MAPK, is known to regulate cell migration in various cells by activating signals such as heat shock protein 27 and LIM-kinase 1 (29, 30). Thus, phosphorylation of MK2 may be a critical factor in regulating RIF-induced morphological changes and the migration of HepG2 cells, although it remains to be proven in future investigations.

CAR (NR1I3) belongs to the same NR1I subfamily as PXR does: CAR and PXR can be activated by an overlapping group of therapeutics and activate the overlapping target genes (31). Similar to PXR, CAR activated the GADD45β gene in HepG2 cells (32). Furthermore, CAR was also found to activate the Gadd45β gene in mouse livers, and this activation occurred independently from the NFκβ-mediated pathway (32, 33). In mouse primary hepatocytes, CAR up-regulated Gadd45β, repressing TNFα-induced phosphorylation of JNK and JNK-mediated cell death (34). In supporting the hypothesis that CAR may regulate cell death via Gadd45β gene, a recent study with Gadd45β KO mice nicely demonstrated that Gadd45β attenuates TNFα-induced JNK phosphorylation and liver regeneration after a partial hepatectomy (35). Given various known overlapping functions of PXR, it is anticipated that further investigations will reveal that PXR may play similar roles as well as yet identified new roles. Using Pxr^+/+ and Pxr^−/− mice, we confirmed that PXR induced Gadd45β and increased phosphorylation level of p38 MAPK in mouse liver in vivo (supplemental Fig. S5). Any physiological/pathophysiological role of induction of GADD45β and activation of p38 MAPK signal pathway by PXR in the liver remains virtually unexplored at the present time. Our present findings may have provided new insights into understanding the molecular mechanisms by which therapeutics causes various cell signal responses such as hepatomegaly, cell proliferation, apoptosis, and migration, thereby affecting human susceptibility to therapeutic exposures (7, 36).

In conclusion, GADD45β as the gene directly activated by PXR presents a novel target for future investigation concerned with the molecular mechanism by which the therapeutic activation of PXR can cause beneficial as well as adverse effects to liver physiology. Because various nuclear hormone receptors are known to cross-talk with the members of the GADD45 family (37), PXR activation by therapeutics and the resulted induction of GADD45β gene may also modulate the nuclear receptor-mediated hormonal responses and homeostasis.