Hepatic oxidative stress activates the Gadd45b gene via degradation of the transcriptional repressor STAT3

Materials

Wy-14,643 and anti-FLAG antibody was purchased from Sigma (St Louis, MO, USA). Customized anti-PPARα antibody was obtained from Affinity Bioreagents (Golden, CO, USA). Anti-STAT3 (for ChIP assay and Western blotting), anti-p21 and anti-GFP antibodies were obtained form Santa Cruz Biotechnology (Santa Cruz, California, USA). Anti-STAT3 (for immunohistology) and anti-pSTAT3 antibodies were purchased from Novus Biologicals (Littleton, CO, USA). Anti-pp65, anti-p65, anti-JNK and anti-pJNK antibodies were obtained from Cell Signaling Technology (Beverly, MA, USA). HRP-conjugated anti-HA antibody was purchased from Roche Applied Science (Indianapolis, IN, USA). Anti-GAPDH antibody was purchased from EMD Millipore (Billerica, MA, USA). JetPEI transfection reagent was purchased from VWR International (Radnor, PA, USA).

Animals

Seven- to eight-week-old male WT and Ppara-null mice (129/Sv background, five animals per group) or Acox1-null mice (C57BL/6 background, three animals per group) or Stat3-floxed (Stat3^F/F) or Stat3-floxed (12) mice harboring the albumin-Cre recombinase gene (Stat3^F/F-AlbCre on a C57BL/6 background) were housed in a pathogen-free animal facility under standard 12 h light/dark cycle with water and chow ad libitum. For Wy-14,643 treatment, mice were fed a normal chow diet (control) or a diet with 0.1% w/w Wy-14,643 (Bioserv, Frenchtown, NJ, USA). The Ikkb-floxed mice (Ikkb^F/F)(13), obtained from Michael Karin, were bred to the albumin-Cre recombinase gene (14) (C57BL/6 background) to obtain Ikkb^F/F-AlbCre mice. All experiments were carried out under the Association for Assessment and Accreditation of Laboratory Animal Care guidelines with approval from the NCI Animal Care and Use Committee.

Cell culture

Hepa1c1c7 cells from ATCC were maintained at 37°C in a humidified atmosphere of 5% CO₂/95% air in DMEM supplemented with 10% heat-inactivated FBS and 50U/ml of penicillin/streptomycin mixture (Invitrogen). Cells were grown to 60–80% confluence and trypsinized with 0.05% trypsin containing 2 mM EDTA.

Plasmids

For mammalian expression, mouse Ppara or Acox1 coding regions were subcloned into the pEGFPc1 vector (Clontech, Mountain Veiw, CA, USA) using XhoI-BamHI or AscI-PacI (extra added restriction sites), respectively. The Stat3c cDNA (provided by Arthur Hurwitz, National Cancer Institute), a dominant positive mutant form for STAT3, was subcloned into the CMV-flag and pmCherry vector using XhoI and BamHI restriction sites. A plasmid encoding HA-ubiquitin was kindly provided by Dong Seok Kim (National Cancer Institute, National Institutes of Health). pHyper-cyto vector (Everogen, Moscow, Russia) was used as an intracellular H₂O₂ sensor.

In vivo DNA delivery

Twenty micrograms of pEGFP-Acox1 or pHyper-cyto vector were dissolved in 2.2 ml of TransIT-QR delivery solution followed by injection through the tail vein in 3 to 5 sec. One day later, the mice were sacrified or fed with Wy-14,643 diet (0.1% w/w) for 24 h. For H₂O₂ measurment, fresh liver pieces (~3 mm³) were squeezed on the slide glass by force using a cover glass. Images were taken under GFP-filtered fluorescence microscopy under the same conditions.

Quantitative real-time PCR (qPCR)

Total RNA was isolated from the frozen/fresh mouse liver or Hepa1c1c7 cells using Trizol (Invitrogen, Carlsbad, CA). Complementary DNA (cDNA) was synthesized from 1 μg of total RNA using a SuperScript II reverse transcriptase kit (Invitrogen, Carlsbad, CA). qPCR was performed using an Applied Biosystems Prism 7900HT Sequence Detection System (Foster City, CA). All reactions were performed in a 10 μl volume consisting of 25 ng cDNA, 300 nM of each primer and 5 μl of SYBER Green PCR Master Mix (Applied Biosystems) at 95°C for 10 min and 40 cycles of 95°C for 3 s and 60°C for 30 s. Expression levels of mRNA were normalized to β-actin as the internal standard. Primers for the qPCR were designed using qPrimerDepot (http://mouseprimerdepot.nci.nih.gov).

Western blotting

Mouse liver or cultured cells were lysed with RIPA lysis buffer (150 mM NaCl, 0.5% Triton X-100, 50 mM Tris–HCl, pH 7.4, 25 mM NaF, 20 mM EGTA, 1 mM DTT, 1 mM Na₃VO₄ and protease inhibitor cocktail) for 30 min on ice, followed by centrifugation at 14,800g for 15 min. Protein concentrations were measured with bicinchoninic acid (BCA) reagent. Protein (30–60 μg) was electrophoresed on a 4–15% gradient tris-HCl gel (Bio-Rad, Hercules, CA) and transferred onto a polyvinylidene difluoride membrane in tris-glycine buffer (pH 8.4) containing 20% methanol. The membrane was blocked with 5% fat-free dry milk in phosphate buffered saline containing 0.1% Tween-20 (PBST) for 1 h. The membranes were probed with primary antibodies and horseradish peroxidase-conjugated secondary antibodies using standard western blotting procedures. Proteins were visualized using the femto signal chemiluminescent substrate (Pierce) under the image analyzer (Alpha Innotech Corp., San Leandro, CA).

Histological and immunohistochemical analysis

For microscopic and macroscopic examination of mouse liver, fresh livers were fixed in 10% buffered formalin and embedded with paraffin. The tissue sections (4 μm) were stained with hematoxylin and eosin (H&E) (Sigma-Aldrich, St. Louis, MO, USA). Anti-STAT3 antibody was use for immunohistochemistry. In situ hybridization for the Gadd45b mRNA in the liver samples were analyzed in the Histoserv, Inc (Germantown, MD)

Chromatin immunoprecipitation (ChIP) Assay

Mouse livers were fixed with 40 ml of 1% formaldehyde by perfusion at 4 ml/min followed by neutralization with same amount of 1 X glycine solution. The livers were chopped and filtered under 40 micron-size mesh. The filtered liver samples were washed with cold PBS and centrifuged at 900g for 2 min at 4°C. Chopped liver:PBS (1:1) mixtures (150 to 200 μl) were lysed with 500 μl of lysis buffer (150 mM Tris-HCl, pH 8.0; 10 mM EDTA; 1% SDS; 300 μM PMSF; protease inhibitor cocktail) and washed with same buffer briefly. Then the samples were sonicated using the Bioruptor system (Cosmo Bio, Tokyo, Japan) for 15 min (30 sec ON/30 sec OFF mode) in an ice-water bath and centrifuged at 12,000 rpm for 5 min at 4°C. Each sample (50 μl) was mixed with 450 μl ChIP dilution buffer (0.01% SDS, 1% Triton X-100, 2 mM EDTA, 150 mM NaCl, 20 mM Tris-HCl, pH.8.0, and protease inhibitor cocktail). The samples were pre-incubated with protein A/G magnetic bead (Thermo Scientific, Rockford, IL, USA) for 30 min and washed with equal amount of ChIP dilution buffer to minimize the nonspecific binding. Then samples were incubated with 6 μg of normal rabbit IgG or anti-STAT3 overnight at 4°C in a rotary shaker. After sample washing, DNA was eluted and reverse-crosslinked with 150 μl of freshly prepared 2 X elution buffer (1% SDS and 0.1 M NaHCO3) containing 6 μl of 5 M NaCl and proteinase K at 65°C for 2 h followed by DNA purification using the Qiagen PCR purification kit. Quantitative real time PCR and conventional PCR were performed using 1 μl of ChIP DNA samples from the 50 μl of purified samples using the following primers for the putative STAT3 binding regions in the Gadd45b promoter region: R1, 5′-AAA GTT CAG GCC AGC CTC TCG-3′ and 5′-AAG TTT CTC TCC TGA TTT CCC GGC-3′; R2, 5′-TCC CTT TCC GGA AGC TGC GTG AG-3′ and 5′-AAT GTC CCT GGC GTG AGA GG-3′; R3, 5′-CGC CTT TAA TCC CAG CAA GCA CTA-3′ and 5′-TGA CTG TCC TGG AAC TCA CTC TGT-3′; R4, 5′-TAT TCC AGG TTG GGC TAT AGC GGT-3′ and 5′-CTG GTG ATG AAT CCA GGA CTT CGT GAT G-3′; R5, 5′-TCC AGA ATC AGT TTG GAT GCC ACC-3′ and 5′-TTT CAT ACT AGG CCA AGC CTA CCC-3′; R6, 5′-TCC TCT GGG TCA AAG CCC AAG TTT-3′ and 5′-TGT CGC CAG AAC AGA GGG AAG AAA-3. The expect PCR sizes were 82 bp, 80 bp, 96 bp, 81 bp, 85 bp and 117 bp, respectively. Conventional PCR was performed using KOD Hot Start DNA polymerase kit (Millipore, Billerica, MA, USA) under the following reaction conditions: initial activation step at 95°C for 2 min, 29 cycles of 95°C for 20 sec, 60°C for 10 sec, 70°C for 20 sec, and a final extension at 70°C for 1 min. DNA binding enrichment were determined using the method as followed; ½^(Ct_IgG −Ct_input): ½^(Ct_STAT3−Ct_input).

Statistical analysis

Experimental values are expressed as mean ± SD. Statistical analysis was performed by two-tailed Student’s t test or Mann Whiney test for unpaired data, with p < 0.05 considered statistically significant.

PPARα agonist Wy-14,643 induced Gadd45b mRNA in the wild-type mice but not in Ppara-null mice

To investigate the role of PPARα in regulation of Gadd45b expression, wild-type (WT) and Ppara-null mice were treated for 24 h with Wy-14,643. Wy-14,643 induced mRNAs from Gadd45b and Acox1, a bonafide PPARα target gene, in the WT mice; no induction was noted in Ppara-null mice (Fig. 1A). In addition, Gadd45b mRNA induction by Wy-14,643 was also confirmed using in situ hybridization (Fig. 1B).

PPARα induction of Gadd45b mRNA is not due to binding to the Gadd45b promoter region

PPARα activation markedly induced Gadd45b mRNA in livers from WT mice but not in Ppara-null mice. However, no PPRE was observed in the Gadd45b promoter region. A series of Gadd45b promoter-luciferase constructs were made (Supporting Fig. 1A) and none of these constructs were activated by Wy-14,643 and PPARα with addition of the obligate dimerization partner RXRα (Suppl. Fig. 1B). Interestingly smaller Gadd45b 5′-flanking region showed higher luciferase activity than the longer constructs (Supporting Fig. 1C) suggesting the presence of repression elements in the Gadd45b promoter. These data suggest that PPARα was not a direct transcription factor controlling Gadd45 induction.

Wy-14,643-induced STAT3 reduction results in induction of Gadd45b mRNA

Gadd45b induction by PPARα could be regulated directly or indirectly. No PPRE’s were detected in the upstream and first exon of the Gadd45b gene and reporter gene assays using a PPARα expression vector and Wy-14,643 did not show direct activation of the Gadd45b promoter. Based on a previously proposed mechanism on Gadd45b regulation (15–16), the expression of NF-κB (p65) and JNK were measured in WT mice and Ppara-null mice after treatment with Wy-14,643. Western blotting showed that there were no significant changes in the phosphorylation of JNK between the groups of WT and WT+Wy-14,643, however, phosphorylation of JNK was reduced in both Ppara-null and Ppara-null +Wy-14,643 groups. In addition, phosphorylation of NF-κB (p65) was markedly decreased preferentially in the Wy-14,643-treated WT mice (Supporting Fig. 2) thus suggesting that NF-κB was not involved in induction by Wy-14,643. Signal transducer and activator of transcription 3 (STAT3) protein was also diminished by Wy-14,643 treatment in WT but not in the Ppara-null mice. The level of p21, a cell cycle inhibitor, was determined because it is under control of GADD45b (17–18). The results showed that p21 was highly induced by Wy-14,643 in WT but not in the Ppara-null mice (Fig. 2A). To further examine the role of STAT3, Stat3 mRNA levels were measured in livers from the WT and Ppara-null mice after Wy-14,643 treatment; Stat3 mRNA was significantly decreased only in the WT and not in Ppara-null mice treated with Wy-14,643 (Fig. 2B). In agreement with the Western blotting data, nuclear STAT3 was significantly decreased only in Wy-14,643-treated WT mice as revealed by immunohistochemistry (Fig. 2C). To assess more directly STAT3 function in the regulation of Gadd45b, Hepa1c1c7 cells were treated for 24 h with different concentrations of VI S3I-201, a STAT3 inhibitor, revealing a robust induction of Gadd45b mRNA (Fig. 2D), inactivation of STAT3 (Fig. 2E) and inhibition of cell proliferation (Fig. 2F) in a dose-dependent manner. In addition, primary hepatocytes isolated from the WT and Ppara-null mice were treated with VI S3I-201 (150 μM) for 24 h, revealing that chemical inhibition of STAT3 induced the Gadd45b mRNA in both WT and Ppara-null primary hepatocytes (Supporting Fig. 3).

Increased expression of Gadd45b in liver-specific Stat3 knockout mice

To support a role for STAT3 in the regulation of Gadd45b, liver samples from Stat3^F/F and Stat3^F/F-AlbCre mice were subjected to qPCR measurement of Gadd45b mRNA. Gadd45b mRNA was markedly elevated in livers of untreated Stat3^F/F-AlbCre (Fig. 3A). Gadd45b induction by Wy-14,643 treatment was also robust in both Stat3^F/F and Stat3^F/F-AlbCre mice (Supporting Fig 4A) thus providing further in vivo evidence that STAT3 may, in part, play a role as a repressor in Gadd45b gene regulation. However, the Stat3^F/F-AlbCre model is of limited value to explain the entire signaling cascade involved in Gadd45b induction, as no significant changes in p65 phosphorylation were noted in between Stat3^F/F or Stat3^F/F-AlbCre mice (Fig. 3B).

Although p-p65 was suppressed by Wy-14,643 treatment suggesting suppression of NF-kB, earlier studies in HeLa cells indicated that NF-kB plays a role in regulation of the human GADD45B promoter (19). To more directly determine a role for NF-kB in regulation of the Gadd45b gene, the Ikk^F/F-AlbCre mouse model (20) was used to determine whether this factor is involved in Gadd45b induction by Wy-14,643 and PPARα in mouse liver. IKKb is an upstream kinase of NF-kB, and absence of this gene suppresses NF-kB. Ikkb^F/F and Ikkb^F/F-AlbCre mice were fed with/without Wy-14,643 (0.1% w/w) diet for 1 day and hepatic Gadd45b mRNA was measured. No difference in Gadd45b mRNA induction by Wy-14,643 was noted between the Ikkb^F/F and Ikkb^F/F-AlbCre mice (Supporting Fig. 4B). These results indicate that NF-kB might not be a critical inducting factor involved in Gadd45b induction by PPARα in mouse liver.

STAT3 binding to the Gadd45b promoter was diminished by Wy-14,643

Based on the results implicating STAT3 in regulation of the Gadd45b gene, potential STAT3 binding sites on the 5′-flanking region of Gadd45b promoter were found using Genomatix software (http://www.genomatix.de). After selection of possible STAT3 binding regions (Fig. 4A), chromatin immunoprecipitation (ChIP) was performed using anti-STAT3 antibody with WT mice fed Wy-14,643 (0.1%, w/w) diet for 1 day. ChIP analysis showed STAT3 binding to several sites in the 5′-flanking region of the Gadd45b promoter; STAT3 was significantly decreased in liver nuclear extracts from Wy-14,643-treated mice (Fig. 4B and 4C).

Gadd45b expression was induced by ACOX1 overexpression in the mouse liver

A potential role for oxidative stress in induction of Gadd45b was investigated since Wy-14,643 treatment is known to elevate peroxisomal ACOX1 that produces H₂O₂ as a byproduct of fatty acid β-oxidation. Forced expression of ACOX1 by hydrodynamic tail vein injection with pEGFP-Acox1 in the liver of WT mice significantly induced Gadd45b mRNA level (Fig. 5A), decreased STAT3 protein levels and inactivated p65 protein (Fig. 5B). Likewise, transient overexpression of ACOX1 also resulted in lower STAT3 protein expression (Fig. 5C). To further verify these results, Acox1-null mice were administered Wy-14,643 for three days and mRNAs encoding Gadd45b and the PPARα target genes such as Acsl1, Cyp4a10 and Acadm were measured. Wy-14,643 treatment had no effect on Gadd45b and other PPARα target gene-encoded mRNAs in the Acox1-null mice (Fig. 5D). Stat3 mRNA (Fig. 5D) and STAT3 protein (Fig. 5E) levels were not changed in Wy-14,643-treated Acox1-null and Acox1-null mice. Hepatic H₂O₂ levels also showed no significant differences between the two groups (Supporting Fig. 5).

H₂O₂ induced Gadd45b and promoted STAT3 inactivation

To more directly determine the role of H₂O₂ in Gadd45b induction, Hepa1c1c7 cells were treated with different concentrations of H₂O₂ for 2 h resulting in robust induction of Gadd45b mRNA (Fig. 6A) and inhibition of phosphorylation of STAT3 (Fig. 6B). There were no significant changes in the NF-κB activation by H₂O₂ in the early time points (Fig. 6B). WT mice were then subjected to hydrodynamic injection with pHyper-cyto (20 μg) plasmid for one day and fed with Wy-14,643 (0.1% w/w) diet for one day to measure intracellular H₂O₂ using the expressed fluorescent sensor protein, HyPer (21). The results revealed that Wy-14,643 treatment increased of H₂O₂ as shown by increased green fluorescence imaging (Fig. 6C). In addition, H₂O₂ levels were directly measured in the livers of Acox1-null and Acox1-null+Wy-14,643 mice, revealing no significant differences in either cytosolic or mitochondrial H₂O₂ was noted between the two groups (Supporting Fig. 5).

ROS, especially H₂O₂, can trigger hypoxia (22). Thus hypoxia-related experiments were carried out because Gadd45b was not induced by Wy-14,643 in cultured primary hepatocytes (Supporting Fig. 6A). Interestingly, Gadd45b mRNA levels from both cobalt chloride (a hypoxia mimic)-induced and 3% oxygen-induced hypoxia conditions were significantly induced (Supporting Fig. 6B and 6C). These results indicate that hypoxia-related factors may be involved in Gadd45b induction.

PPARα activation promotes STAT3 ubiquitination

To investigate whether STAT3 degradation was directly the result of Wy-14,643 PPARα activation, the localization of PPARα and STAT3 was examined. Hepa1c1c7 cells were co-transfected with pEGFP-Ppara and pmCherry-Stat3c constructs for 24 h followed by fluorescence imaging. EGFP-PPARα and STAT3C were co-localized in the nuclear body (Fig. 7A), regarded as the site for ubiquitin-like post-translation modification (23). A binding assay between EGFP-PPARα and STAT3C was carried out by co-immunoprecipitation (co-IP) with nuclear extracts from Hepa1c1c7 cells. STAT3C was bound directly to the EGFP-PPARα protein and the binding was decreased by Wy-14,643 treatment (Fig. 7B). To explore the mechanism of STAT3 degradation, Hepa1c1c7 cells were co-transfected with plasmids expressing EGFP-PPARα and STAT3C-FLAG in the presence or absence of ubiquitin-HA expression for 24 h and then treated with Wy-14,643 for another 24 h. MG132, a proteasome inhibitor, was treated for 2.5 h prior to cell harvest and nuclear extracts subjected to co-IP using anti-FLAG antibody and blotted against HRP conjugated anti-HA antibody. STAT3C underwent ubiquitin degradation in the presence of Wy-14,643 and its degradation was intensified by MG132 (Fig. 7C).