INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Heme oxygenase (HO,¹ EC 1.14.99.3) is the rate-limiting enzyme of heme catabolism. It catalyzes the breakdown of heme into equimolar amounts of carbon monoxide, iron, and biliverdin. Two isoforms transcribed from separate genes have been characterized; HO-1, a ubiquitous, inducible form found in large quantities in liver and spleen; and HO-2, a constitutively active form found mainly in the brain and testes (1-3). Many chemical and environmental stimuli are known to induce HO-1, including its substrate heme (4-6), other metalloporphyrins (4, 7-9), transition metals (6, 10, 11), ultraviolet light (12, 13), phorbol esters (14), heat shock (5, 15, 16), and other chemical initiators of cellular stress responses, such as hydrogen peroxide (12), lipopolysaccharide (17, 18), and arsenite (12, 15, 19). Recent studies by Poss and Tonegawa (20, 21) in HO-1-deficient mice have highlighted the important metabolic and cytoprotective roles of this gene. The mice exhibited an incapacity to modulate body iron stores properly and were less resistant to hepatic injury by iron, indicating that HO-1 plays an important role in iron utilization (20). After exposure to oxidative damage-causing agents, such as hemin, hydrogen peroxide, or cadmium, the mice were hypersensitive to cytotoxicity when given additional hemin or hydrogen peroxide. When subsequently challenged with endotoxin, HO-1-deficient mice were highly susceptible to hepatic necrosis or death (21).

Many studies of HO-1 regulation and hepatic heme metabolism have been performed using chick embryo liver cells as the model system (6, 8, 22-30). We have recently provided evidence that supports the use of LMH cells, a chicken hepatoma cell line, as a useful and reliable model system for studying the mechanism of HO-1 induction (5). LMH cells are readily available, transfect more efficiently, and are more homogeneous than chick embryo liver cells.

HO-1 has been characterized as a heat shock protein, and growing evidence supports a role for HO-1 in protecting cells from oxidant stress (12, 13, 19, 31-34). Several cis-acting promoter elements involved in mediating HO-1 gene expression have been elucidated (35-38). In one study (7), stably transfected HO-1 reporter gene constructs were used to locate elements required for induction of the murine HO-1 gene by heme and heavy metals. Putative regulatory elements were identified between 3.5 kb and 12.5 kb upstream from the transcription start site. Recently, evidence for a basal level inducer and heme-responsive element located as far as 10.5 kb upstream of the transcription start site in the mouse HO-1 gene promoter was presented (14). Other work has located regions of the HO-1 promoter that mediate cadmium, heat shock, hypoxia, and lipopolysaccharide (LPS) responsiveness (14, 16, 17, 38-41). However, the signal transduction pathways and transcription factor complexes that target these elements have been largely unexplored.

Our laboratory has shown that induction of HO-1 gene expression by heme is mediated through a pathway distinct from that mediated by transition metals (6, 9, 42, 43). Distinctions between the mechanisms utilized by metals, heme, and heat shock to induce HO-1 have also been described (4, 15, 16, 18, 44). Despite these differences, the effects of diverse factors on HO-1 gene expression appear to be regulated at the transcriptional level, suggesting that multiple signal transduction pathways mediate induction of HO-1 gene transcription in response to a multitude of cellular stimuli.

The mitogen-activated protein (MAP) kinases are serine/threonine protein kinases that have been shown to be activated under conditions similar to those that induce HO-1 transcription, e.g. following exposure of cells to phorbol esters, cytokines, ultraviolet light, heat shock, LPS, ceramide, and inducers of oxidative stress (44-52). Three major MAP kinase subfamilies that mediate physiological responses have been described: extracellular-regulated kinase (ERK), c-Jun N-terminal kinase (JNK), and p38 (a homolog of the yeast HOG1 kinase). MAP kinases are components of signaling cascades which, in response to extracellular stimuli, target transcription factors, resulting in the modulation of gene expression. The MAP kinase pathway leading to activation of ERK has been studied extensively. Cell surface receptor tyrosine kinases activate a signaling cascade involving Ras, Raf, MEK (MAP kinase/ERK kinase), and ERK, which then targets other kinases or transcription factors (53-58).

Arsenite has been shown recently to activate MAP kinases (44, 45, 59-62); however, this arsenite-mediated activation has not been linked to a cellular gene response. Arsenite is also proposed to affect gene expression by modulating the activities of transcription factor complexes bound to AP-1 elements in the promoter regions of several genes (60, 63, 64). Putative AP-1 sites are present in the promoter regions of mammalian (14, 17, 35, 38) and chicken HO-1 (65). Since the MAP kinase signaling cascades have been shown to target AP-1 elements, they are potential candidates as mediators of the arsenite induction of HO-1 (55, 66).

In this study, we investigated the ability of arsenite to increase transcription of endogenous HO-1, and the activity of transfected luciferase reporter gene constructs under control of the HO-1 promoter. Transient transfection assays were used to investigate the mechanism of arsenite-mediated HO-1 gene expression. In LMH cells, the activities of MAP kinases, ERK, JNK, and p38 were increased by treatment with arsenite. Activation of MAP kinases correlated with arsenite-mediated induction of endogenous HO-1 mRNA expression. Activated components of the ERK and p38 MAP kinase signaling pathways increased gene expression from an HO-1 promoter-driven luciferase reporter gene construct. A p38 inhibitor, a MEK inhibitor, and dominant negative components of the ERK and p38 MAP kinase pathways were able to block most of the arsenite-mediated induction of HO-1. In contrast, for the JNK pathway, activated components were unable to induce HO-1 gene expression and dominant negative components were unable to block arsenite induction. HO-1 reporter constructs containing a mutated AP-1 element were unresponsive to arsenite treatment. These experiments implicate a role for the ERK and p38 MAP kinase families, and at least one AP-1 element, in the sodium arsenite-mediated induction of HO-1 gene expression.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Materials-- Tissue culture dishes were from Corning, Corning, NY; culture flasks were from Falcon, VWR Scientific, Bridgeport, NJ. Heme was from Porphyrin Products, Logan, UT. Chloroform, isopropanol, and sodium vanadate were from Fisher. Actinomycin D, adenosine triphosphate (ATP), -mercaptoethanol, bovine serum albumin, cadmium chloride, cobalt chloride, dimethyl sulfoxide, EDTA, EGTA, formaldehyde (37%, v/v), formamide, glycylglycine, O-nitrophenyl--galactopyranoside, penicillin/streptomycin, phenylmethylsulfonyl fluoride, piperacillin, sodium arsenite, sodium dodecyl sulfate, Triton X-100, and trypsin were from Sigma. Dithiothreitol (DTT) was from Aldrich. Kinase inhibitors PD98059 and SB203580 were from Calbiochem. Gelatin and fetal bovine serum were from Difco. Waymouth's MB 752/1 media, Opti-MEM, and LipoFECTAMINE^® reagent were from Life Technologies, Inc. Ultraspec RNAzol^TM was from Biotecx, Houston, TX. Nitrocellulose (0.45 mm) was from Schleicher & Schuell. All ³²P-radionucleotides were from NEN Life Science Products. LMH cells and the pGAD-28 plasmid were generous gifts from D. L. Williams (Department of Pharmacological Sciences, SUNY, Stony Brook, NY). The  FIX II clone, containing genomic chicken heme oxygenase-1 promoter sequence, and pCHO3.6-CAT were provided by T. H. Lu (Division of Digestive Disease and Nutrition, University of Massachusetts Medical Center, Worcester, MA). The pGL3 Basic and pGL3 Control plasmids were gifts from G. Gil (Division of Digestive Disease and Nutrition, University of Massachusetts Medical Center, Worcester, MA). The pPGK-gal plasmid was a gift from P. Dobner (Department of Molecular Genetics and Microbiology, University of Massachusetts Medical Center, Worcester, MA). Expression vectors for signaling pathway components and MAP kinases have been described: ERK2 (67), MEK1 (68), MEKK1 (69, 70), JNK1 (71), MLK3 (72), MKK6 (73), and Ras and Raf (74). The expression vector for dominant negative c-Jun (TAM67) has been described (75). The primers GLprimer2 and RVprimer3, Wizard^TM plasmid DNA preparation kits, and Luciferase assay reagent were purchased from Promega, Madison, WI. The QuickChange^TM site-directed mutagenesis kit was from Stratagene, La Jolla, CA. DNA sequencing was performed by Dana Farber Cancer Institute, Boston, MA, and by the Nucleic Acids Facility, University of Massachusetts Medical Center, Worcester, MA. All chemicals were of the highest purity available.

Cell Culture and Treatments-- LMH cells were maintained in Waymouth's MB 752/1 medium supplemented with 100 units/ml penicillin, 100 µg/ml streptomycin, and 10% (v/v) fetal bovine serum, and were routinely passaged twice a week (76). For some experiments, 1 µg/ml piperacillin was added to the culture medium. Sodium arsenite, cadmium chloride, cobalt chloride, and heme were prepared as described (6, 8, 22, 24-26, 76, 77). Hydrogen peroxide and LPS were freshly prepared on the day of treatment and added directly to the culture medium.

Transfections-- Cells were plated in six-well plates coated with 0.1% gelatin at a density of 3.2 × 10⁵ cells/well. For transfections, each well received 0.5 µg of pPGK-gal, and 0.5-1.0 µg of cHO-1 promoter/reporter and/or MAP kinase plasmid DNA using LipoFECTAMINE^® (0.25 µg of DNA/µl of reagent), according to the manufacturer's protocol. Total DNA transfected was kept constant by adding pBLUESCRIPT KS II+ plasmid DNA. Incubation was continued overnight at 37 °C, 5% CO₂ for a total of 20-24 h. Cells were incubated with serum-free Waymouth's medium for at least 20 h prior to treatment with selected chemicals or harvest.

Harvest, Isolation, and Quantitation of mRNA-- The cells from one 6-cm culture dish were harvested directly into 0.5 ml of Ultraspec RNAzol^TM and total RNA isolated as described (8). Purity was assessed by determining the absorbance ratio at 260 nm/280 nm (78). RNA concentrations were estimated from the absorbance at 260 nm (1 absorbance unit = 40 µg/ml RNA) (79). RNA samples were prepared and loaded onto dot blots essentially as described (8). Radiolabeled probes were generated using the Ready-To-Go^TM DNA labeling beads from Amersham Pharmacia Biotech. The amount of specific mRNA was normalized to glyceraldehyde-3-phosphate dehydrogenase mRNA, measured on a duplicate blot hybridized with a radiolabeled probe, as described previously (8). Templates for probes were polymerase chain reaction products, made using T7 and T3 primers and the cHO-1 cDNA; and the linearized pGAD-28 plasmid with the primers GAP1: 5'-GAA AGT CGG AGT CAA CGG ATT TG-3' and GAP2: 5'-TGG CAT GGA CAG TGG TCA TAA GAC-3' for glyceraldehyde-3-phosphate dehydrogenase. Hybridization was performed with probes previously shown by Northern blot to specifically bind only the mRNA of interest. Quantitation of specific binding was performed with the aid of a PhosphorImager, Molecular Dynamics, Sunnyvale, CA.

Assessment of Reporter Gene Activity-- Reporter gene expression and activation was assessed by quantitation of luciferase activity, normalized to -galactosidase activity, and protein content. Experiments in which dominant negative ERK kinase components were transfected were normalized to protein content only as the -galactosidase values were below the linear range. For luciferase activities, transfected cells were washed twice with 1× phosphate-buffered saline, and harvested by scraping in 250 µl of glycylglycine harvest buffer (25 mM glycylglycine, pH 7.8, 15 mM magnesium sulfate, 4 mM EGTA, 1 mM DTT). Cells were lysed by three cycles of freeze-thaw (3 min in liquid nitrogen followed by 3 min at 37 °C), followed by a 10-min centrifugation at 14,000 × g at 4 °C. The supernatant was retained, and 15-µl aliquots of cell lysate were used for each assay. Luciferase activity measurements were carried out using a Monolight 2010^TM luminometer (Analytical Luminescence Laboratories, Ann Arbor, MI). Relative light units produced in 10 s were recorded and normalized with -galactosidase activities and protein content. For -galactosidase activities, 15 µl of cell lysate was added to 200 µl of Z buffer (60 mM Na₂HPO₄·7H₂O, 40 mM NaH₂PO₄·H₂O, 10 mM potassium chloride, 1 mM magnesium sulfate, 50 mM -mercaptoethanol, adjusted to pH 7.0) and 100 µl of 5 mg/ml O-nitrophenylgalactoside dissolved in 0.1 M potassium phosphate, pH 7.0. The samples were mixed and incubated at 37 °C for 1.5-2 h. The absorbance at 420 nm was measured and used for normalization of luciferase activities (79). Protein concentrations were determined from absorbance at 562 nm measured by the bicinchoninic acid method on a Spectronic GENESYS 2 spectrophotometer, using bovine serum albumin as standard (80).

Immune Complex Kinase Assays-- Immune complex kinase activity assays were performed as described (81). Cells were harvested by scraping into 1.0 ml of 1× Triton lysis buffer (TLB; 20 mM Tris, pH 7.4, 137 mM NaCl, 25 mM -glycerophosphate, 2 mM sodium pyrophosphate, 2 mM EDTA, 1 mM sodium vanadate, 1% Triton X-100, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 5 µg/ml leupeptin, 5 µg/ml aprotinin, 2 mM benzamidine, 0.5 mM DTT). Lysates were centrifuged at 14,000 × g at 4 °C for 15 min. Anti-ERK2, anti-JNK1 (Santa Cruz Biotechnology, Inc.), or anti-p38 (81) antibodies were bound to Protein A-Sepharose beads (10 µl/assay) for at least 30 min and washed twice with 1× TLB prior to adding cell lysates. Pre-bound antibodies were incubated with 300 µl of cell extract in a final volume of 500 µl with 1× TLB. After mixing at 4 °C for at least 3 h, the immunoprecipitates were washed three times with 1× TLB, and once with 1× kinase assay buffer (25 mM HEPES, pH 7.4, 25 mM -glycerophosphate-Na⁺, 25 mM magnesium chloride, 0.1 mM sodium vanadate, 0.5 mM DTT). After aspiration, the kinase assay was set up with 10 µl of Protein A-Sepharose beads/antibody/kinase complex, 26 µl kinase assay buffer, 2 µl of kinase substrate (2 µg of GST-Elk1 for ERK (70), GST-ATF2 for p38 (81), and GST-c-Jun for JNK (71)), 1 µl of 1 mM ATP, 1 µl of carrier-free [-³²P]ATP (approximately 10 µCi/µl) for a final volume of 40 µl. Samples were incubated at room temperature for 30-45 min. The reactions were stopped with 2× SDS-polyacrylamide gel electrophoresis sample buffer, and 15-20 µl were run on SDS-polyacrylamide gels. Results were visualized by autoradiography and quantitated using a PhosphorImager and ImageQuant software.

Subcloning-- The pCHO3.6-Luc reporter plasmid was constructed by subcloning 3728 base pairs of the chicken HO-1 proximal promoter into the pGL3 Basic plasmid vector upstream of the luciferase reporter gene. The ligation junctions were verified by sequencing. A FIX^TM II vector containing approximately 12 kilobases (kb) of genomic chicken heme oxygenase-1 (cHO-1) sequence including some coding sequence and ~10.5 kb of promoter sequence was digested with XbaI and XhoI. The 7.1-kb fragment was ligated into a pGL3 Basic luciferase reporter vector that had been digested with NheI and XhoI, and transformed into DH5 cells. Colonies that were positive for pCHO7.1-Luc were verified by sequencing the junctions using the commercial primers GLprimer2 and RVprimer3 from Promega. Deletions were constructed by digesting pCHO7.1-Luc with MluI and the respective enzyme for each construct. When enzymes left a 3' overhang, the DNA was treated with T4 polymerase; when a 5' overhang was left, the DNA was treated with Klenow to blunt the ends. Following ethanol precipitation, blunt-end ligation was performed. The corresponding restriction enzyme site was ligated with the MluI site at 7085 base pairs. Positive colonies were identified by DNA isolation, followed by a diagnostic XbaI and XhoI restriction enzyme digest. The ligation junctions were verified by sequencing.

Site-directed Mutagenesis-- For pCHO2.5mut1-Luc and pCHO2.5mut2-Luc, the AP-1 element located 1576 base pairs from the transcription start site in pCHO2.5-Luc was mutated using the QuickChange^TM kit and the following polymerase chain reaction primers: 5'-GCAGAGCAAG ACAGGAAAAG CATGGCTTCG TCAGGCTGGG AGCGCTGAG-3' and 5'-CTCAGCGCTC CCAGCCTGAC GAAGCCATGC TTTTCCTGTC TTGCTCTGC-3' for mutant 1, and 5'-GACAGGAAAA GCATGGCGGA GTCGGGCTGG GAGCGCTGAG-3' and 5'-CTCAGCGCTC CCAGCCCGAC TCCGCCATGC TTTTCCTGTC-3' for mutant 2.

Statistical Analysis of Data-- Experiments were repeated two to four times; except for immune complex kinase assays, every experiment included at least triplicate samples for each treatment group. Representative results from single experiments are presented. Statistical analyses were performed with JMP 3.0.2 software (SAS Institute, Cary, NC). Differences in mean values were assessed by analysis of variance, with the Tukey-Kramer correction for multiple pairwise comparisons, or Dunnett's test versus a control. For experiments with non-normally distributed data, the Wilcoxon/Kruskal-Wallis (rank-sum) test was used. p values <0.05 were considered significant.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Induction of Transfected Heme Oxygenase-1 Promoter Luciferase Reporter Constructs by Sodium Arsenite-- Sodium arsenite is a potent inducer of HO-1 mRNA in chick embryo liver cells (12, 19, 43, 82). In LMH cells, 75 µM sodium arsenite gave a peak induction of HO-1 mRNA (3.9-fold) at 4 h (data not shown). The highly reproducible, robust induction of HO-1 mRNA expression by arsenite provides a reliable system to study the signaling mechanisms involved. To aid this study, several HO-1 promoter-luciferase reporter constructs were made. The longest of these constructs included 7.1 kb of the chicken HO-1 promoter cloned upstream of the firefly luciferase reporter gene (Fig. 1). The HO-1 promoter-luciferase reporter constructs were then tested for responsiveness to sodium arsenite to define regions of the promoter that may be responsible for the induction effects. Although arsenite induced several constructs, the highest -fold of induction was observed for pCHO5.6-Luc and pCHO7.1-Luc (Fig. 1).

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Sodium arsenite induction of heme oxygenase-1 promoter-luciferase reporter constructs. Restriction enzyme sites used for creating reporter constructs are shown in relation to the chick heme oxygenase-1 TATA box (3' end = 26 base pairs), txn = transcription start site (+1), and the XhoI cloning site (+54). Abbreviated restriction enzyme recognition sites are designated as follows: F, PflMI; P, PmlI; N, NheI; B, BglII; E, EcoRI; T, Tth111I. The restriction enzyme and sequence location used to create each construct are identified beside each representation. Sequence of the proximal promoter region has been deposited with GenBank (accession number U95209). The pCHO3.6-Luc construct was subcloned into pGL3 by ligation of a fragment taken from a chloramphenicol acetyltransferase reporter construct. See "Experimental Procedures" for details of creation of constructs. For ARS induction, LMH cells were transfected with plasmid DNA using LipoFECTAMINE®. Following transfection, some cells were treated with 75 µM ARS for 6 h. Cells were harvested by scraping, lysed by freeze/thaw, and assays performed on 15 µl of lysate. Luciferase activities (relative light units) were normalized to -galactosidase activities and protein content. Data are presented as mean ± S.E., n = 3. *, differs from no treatment and constructs containing less than 3.6 kilobases of HO-1 promoter, p < 0.001. , differs from 5.6- and 7.1-kilobase constructs, p < 0.01.

Responsiveness of pCHO7.1-Luc to Inducers of Cellular Stress-- To determine the usefulness of this construct as a tool for studying HO-1 gene expression under conditions of cellular stress, we tested the effects of several chemicals implicated in stimulating cellular stress responses. The largest construct, pCHO7.1-Luc, was used since it produced the most consistent induction when treated with arsenite. Sodium arsenite had the greatest effect on reporter gene expression, with a 4.3-fold increase in normalized luciferase activity (data not shown). Lesser degrees of induction were observed with this construct after exposure to hydrogen peroxide (1.6-fold) and cadmium chloride (2.3-fold). LPS and cobalt chloride have been shown to induce HO-1 in some experimental systems; however, no induction was observed for LPS or cobalt chloride with pCHO7.1-Luc (data not shown).

Optimization of Arsenite-mediated Induction of pCHO7.1-Luc-- Further analysis was performed on the arsenite-induced increase in reporter gene expression in cells transfected with pCHO7.1-Luc. Dose-response and time-course experiments established that the most effective dose for arsenite was 75 µM and peak induction occurred at 6 h (Fig. 2). Therefore, these conditions were used with this construct to further investigate the mechanism of sodium arsenite induction of HO-1 gene transcription.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Dose and time dependence of induction of pCHO7.1-Luc by sodium arsenite. LMH cells were transfected with plasmid DNA using LipoFECTAMINE®. Following transfection, cells were treated with increasing concentrations of ARS for 8 h (panel A) or 75 µM ARS for increasing lengths of time (panel B). Cells were harvested by scraping, lysed by freeze/thaw, and assays performed on 15 µl of lysate. Luciferase activities (relative light units, RLU) were normalized to -galactosidase activities and protein content. Data are presented as mean + S.E., n = 3. *, differs from control (no treatment), p < 0.001.

Arsenite Activates MAP Kinases in LMH Cells-- Sodium arsenite acts through a cellular stress mechanism involving the oxidation of cellular sulfhydryl-containing proteins (12, 14, 15, 19, 63, 82, 83). Recent reports have suggested that arsenite activates MAP kinase signal transduction pathways (44, 45, 59-62). Interestingly, many of the inducers of endogenous HO-1 gene expression are also activators of MAP kinases (UV radiation, LPS, growth factors, phorbol esters, heat shock, etc.). Since MAP kinase activities have been shown to be differentially regulated in a cell-type specific manner, we assessed the ability of sodium arsenite and several other known inducers of HO-1, to activate the MAP kinases, ERK, JNK, and p38 in LMH cells. Cells were treated with 20 µM heme, a known inducer of endogenous HO-1; 75 µM sodium arsenite, a potent inducer of both endogenous HO-1 and HO-1 reporter gene constructs; 1.5 µM cadmium chloride, a known metal inducer of HO-1; or exposed to heat shock (at 43 °C) for 0, 30, or 60 min. Immune complex kinase assays were performed to detect changes in MAP kinase activity (Fig. 3). Treatment with sodium arsenite increased the activity of all three MAP kinases. In contrast, cadmium and heat shock caused slight increases in ERK and p38 activities, and heme had no effect on any of the MAP kinases in LMH cells. To further characterize the arsenite-mediated activation of the MAP kinases, we performed a more detailed time course (Fig. 4). The peak activation times observed were 10 min for ERK, 20 min for p38, and 45 min for JNK.

View larger version (35K): [in this window] [in a new window]   Fig. 3.   Effects of heme, sodium arsenite, cadmium chloride, or heat shock on MAP kinase activity in LMH cells. LMH cells were treated with 20 µM heme, 75 µM sodium arsenite, 1.5 µM cadmium chloride, or exposed at 43 °C (heat shock) for 0, 30, or 60 min prior to harvest. Panel A, immune complex kinase assays were performed for ERK (top), JNK (middle), and p38 (bottom). Bands indicate phosphorylated substrates: GST-Elk1 for ERK, GST-c-Jun for JNK, and GST-ATF2 for p38. Panels B-D, data from immune complex kinase assays were quantitated with the aid of a PhosphorImager and plotted as time courses of ERK, JNK, or p38 MAP kinase activation.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Time course of MAP kinase activation by ARS. LMH cells were treated with 75 µM ARS for the indicated times. Cell lysates were harvested and immune complex kinase assays specific for ERK, JNK, and p38 MAP kinases were performed. Data are plotted as -fold induction relative to untreated control values.

Induction of pCHO7.1-Luc by Components of MAP Kinase Signaling Cascades-- To link activation of MAP kinases with the induction of HO-1 gene expression, several expression constructs encoding constitutively activated or dominant negative components of the MAP kinase signaling pathways were co-transfected with the pCHO7.1-Luc reporter construct (Fig. 5). The constitutively activated MAP kinase components would be expected to increase reporter gene activity if the downstream MAP kinases are involved in transducing cellular signals that control HO-1 gene expression. Constitutively activated forms of the kinase immediately upstream of ERK, MEK1; a kinase proposed to act upstream of JNK, MEKK1; and a kinase upstream of p38, MKK6; were tested for their abilities to induce luciferase gene expression from pCHO7.1-Luc. Activated MEK1 and MKK6 were able to induce reporter gene expression, while no induction was observed for MEKK1 (Fig. 5). As expected, all dominant negative components failed to affect basal level luciferase expression from pCHO7.1-Luc.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Effect of selected MAP kinase pathway components on luciferase reporter gene activity in LMH cells co-transfected with pCHO7.1-Luc. Constitutive activators of ERK MAP kinase (MEK1 S218E-S222D and MEK1 N3-S218E-S222D), JNK MAP kinase (MEKK1), and p38 MAP kinase (MKK6 S207E-T211E) were tested for ability to induce reporter gene activity. Dominant negatives of the ERK pathway (MEK1 K97A), JNK pathway (MEKK1 K432A), and p38 pathway (MKK6 S207A-T211A) were also tested. LMH cells were co-transfected with pCHO7.1-Luc, pPGK-gal, and the indicated MAP kinase construct using LipoFECTAMINE®. Luciferase activities were normalized to -galactosidase activities and protein content. Data are presented as mean + S.E., n = 3. *, differs from control (pCHO7.1-Luc only), p < 0.005. RLU, relative light units.

Involvement of the ERK Signaling Cascade in Arsenite Induction of HO-1-- To investigate the role of ERK in arsenite induction of HO-1, several ERK pathway components (Ras activated and negative, Raf wild type and activated, MEK1 activated and negative, and wild type ERK2) were co-transfected with pCHO7.1-Luc, then left untreated or treated with arsenite. If ERK was important for arsenite signaling, an activated component of the pathway would increase luciferase gene activity in the absence of arsenite, while the negative would block the ability of arsenite to induce luciferase activity from the reporter gene. As shown in Fig. 6, activated Ras, Raf, and MEK1 all increased luciferase gene activity, indicating each component led to induction of HO-1 gene expression. Dominant negative Ras and MEK1 also abrogated induction by arsenite treatment. The wild type ERK2 and wild type Raf gave results similar to those for the pCHO7.1-Luc only control.

View larger version (26K): [in this window] [in a new window]   Fig. 6.   Effects of ERK signaling pathway components on reporter gene activity in LMH cells co-transfected with pCHO7.1-Luc. Wild type Raf, ERK2, or constitutive activators of Ras (Ras 61L), Raf (Raf BXB), and the ERK pathway (MEK1 S218E-S222D) were tested for ability to induce reporter gene activity. Dominant negatives of Ras (Ras 17N) and the ERK pathway (MEK1 K97A) were tested for ability to block arsenite induction of the reporter gene. LMH cells were co-transfected with pCHO7.1-Luc, pPGK-gal, and the indicated MAP kinase construct using LipoFECTAMINE®. Luciferase activities were normalized to protein content. Data are presented as mean + S.E., n = 3. *, differs from untreated, p < 0.01. , differs from untreated control (pCHO7.1-Luc only), p < 0.01. §, differs from arsenite-treated control, p < 0.005. RLU, relative light units.

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Effects of MEK inhibitor, PD98059, on ERK activity and ARS-mediated induction of heme oxygenase-1. LMH cells were left untreated or treated with the indicated concentrations of MEK inhibitor. Some samples received 75 µM ARS 30 min after addition of inhibitor. Panel A, cells were harvested 10 min after addition of ARS (40 min after inhibitor) and ERK activity was measured. Panel B, prior to treatment, cells were transfected with pCHO7.1-Luc, and pPGK-gal. Cells were harvested 6 h after addition of ARS (6.5 h after inhibitor). Luciferase activities were normalized to -galactosidase activities and protein content. Data are presented as mean + S.E., n = 3. **, differs from ARS only control, p < 0.005. *, differs from ARS only control, p < 0.001. RLU, relative light units.

JNK MAP Kinase Is Not Implicated in Arsenite Induction of HO-1-- JNK involvement in HO-1 induction by arsenite was studied by co-transfecting components of the JNK pathway with pCHO7.1-Luc. If JNK were an intermediate in the pathway to HO-1 induction, wild type constructs would either increase luciferase reporter activity or give results similar to control, while dominant negative components would block the arsenite increase in reporter gene activity. Fig. 8 shows that wild type JNK, MEKK1, and MLK3 resulted in luciferase gene activity levels similar to control. However, contrary to the results for ERK components, arsenite treatment continued to increase luciferase gene activity in the presence of co-transfected dominant negative MEKK1 and MLK3.

View larger version (28K): [in this window] [in a new window]   Fig. 8.   Effects of JNK signaling pathway components on reporter gene activity in LMH cells co-transfected with pCHO7.1-Luc. Wild type JNK, and activators of the JNK kinase pathway (MEKK1 and MLK3) were tested for their ability to induce reporter gene expression. Dominant negative JNK MAP kinase pathway components (MEKK1 K432A and MLK3 K144R) were tested for their ability to block arsenite induction of the reporter gene. LMH cells were co-transfected with pCHO7.1-Luc, pPGK-gal, and the indicated MAP kinase construct using LipoFECTAMINE®. Luciferase activities were normalized to -galactosidase activities and protein content. Data are presented as mean + S.E., n = 3. *, differs from untreated control, p < 0.05. RLU, relative light units.

Involvement of p38 MAP Kinase in Arsenite Induction of the HO-1 Gene-- In Fig. 5, an ERK component (MEK) and a p38 component (MKK6) demonstrated the ability to induce luciferase reporter activity. To further delineate a role for p38 in arsenite signaling, a p38 kinase specific inhibitor (SB203580) was tested for the ability to block arsenite induction of HO-1 gene expression. Fig. 9A shows the effects of the p38 inhibitor on arsenite-induced HO-1 mRNA levels in LMH cells. When cells were pretreated with the p38 inhibitor 30 min prior to arsenite treatment, the ability of arsenite to increase HO-1 mRNA levels was decreased to only 47% of the control (no inhibitor). The inhibitor alone caused no significant change in HO-1 mRNA levels.

View larger version (22K): [in this window] [in a new window]   Fig. 9.   Effects of inhibitors of MAP kinase pathways on heme oxygenase-1 gene expression in LMH cells. Panel A, LMH cells were treated with the indicated concentrations of p38 kinase inhibitor (SB203580) alone, or 30 min prior to addition of 75 µM ARS. Cells were harvested 4 h after addition of ARS (4.5 h after inhibitor). HO-1 mRNA was quantitated and normalized to glyceraldehyde-3-phosphate dehydrogenase. Data are presented as mean + S.E., n = 3. *, differs from control (ARS only), p < 0.01. Panel B, prior to treatment, cells were transfected with pCHO7.1-Luc, and pPGK-gal. After transfection, cells were treated with either 30 µM MEK inhibitor (PD98059), or 20 µM p38 inhibitor (SB203580), or both, 30 min prior to addition of 75 µM ARS. Cell lysates were harvested 6 h after addition of ARS (6.5 h after inhibitor(s)). Luciferase activities were normalized to -galactosidase activities and protein content. Data are presented as mean + S.E., n = 3. *, differs from no ARS, p < 0.001. **, differs from no ARS, p < 0.005. , differs from ARS control, p < 0.01. ¥, differs from ARS control, p < 0.05. §, differs from PD + ARS, p < 0.01. RLU, relative light units.

Involvement of an AP-1 Element in HO-1 Gene Expression-- We investigated the role of an AP-1 site as a transcriptional element that may modulate HO-1 gene expression in response to arsenite. We chose the reporter construct pCHO2.5-Luc, which contains a single consensus AP-1 element located at 1576 base pairs, upstream of the transcription start site. The role of this AP-1 site in transcriptional activation was studied by making site-directed mutations in 2 out of 7 base pairs, as illustrated in Fig. 10A. Two separate mutants were made and tested for their ability to be induced by treatment with arsenite (Fig. 10B). Both mutant 1 and mutant 2 were incapable of being induced by arsenite treatment. Similar results were previously obtained in primary chick embryo liver cells by T. H. Lu using a CAT construct containing a different mutation at the same site.² In addition, co-transfection of the wild type pCHO2.5-Luc with a dominant negative c-Jun also blocked the arsenite induction of luciferase reporter gene activity. Together, these data strongly implicate AP-1 as one of the transcription factors that contribute to arsenite-induced HO-1 gene expression.

View larger version (32K): [in this window] [in a new window]   Fig. 10.   Involvement of an AP-1 element in ARS-mediated heme oxygenase-1 expression. Panel A, the pCHO2.5-Luc HO-1 promoter-luciferase reporter construct contains an AP-1 site at 1576 base pairs. Site-directed mutants 1 and 2 contain different two-base changes, indicated in bold and underlining. Abbreviated recognition sites are designated as follows: T, Tth111I; N, NheI; P, PmlI; E, EcoRI. Panel B, LMH cells were transfected with wild type alone or with TAM67 (dominant negative c-Jun), or with mutant 1 or mutant 2, and pPGK-gal. Following transfection, cells were left untreated or treated with 75 µM ARS for 6 h. Luciferase activities were normalized to -galactosidase activities and protein content. Data are presented as mean -fold induction versus untreated control + S.E., n = 3. *, differs from wild type, p = 0.0335.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

The major findings of this study are: 1) sodium arsenite is a potent inducer of both endogenous HO-1 and transfected HO-1 promoter-reporter constructs (Figs. 1 and 2); 2) sodium arsenite activates the MAP kinases ERK, JNK, and p38 in LMH cells (Figs. 3 and 4); 3) activated components of the ERK and p38 MAP kinase pathways are capable of inducing HO-1 gene expression, while dominant negative components block arsenite induction of HO-1 (Figs. 5 and 6); 4) inhibitors of MEK and p38 differentially inhibit the arsenite induction of HO-1 (Figs. 7 and 9); and 5) an AP-1 element at 1576 base pairs is involved in transcriptional activation of HO-1 by arsenite (Fig. 10).

Although there is now much evidence that HO-1 can be induced by various chemical and physical agents, little is known about the signaling mechanisms utilized by these stimuli. Our laboratory previously presented evidence that at least two separate mechanisms of induction exist, one that is heme-dependent, and another (mediated by transition metals) that is heme-independent (6), but the exact mechanism of signal transduction from the cell surface through second messengers to transcription factors, and finally to promoter elements was not delineated. The recent cloning and characterization of the genomic chicken HO-1 gene (65) has allowed us to develop tools for elucidating the mechanisms involved in modulating heme oxygenase-1 gene expression.

Arsenite potently induced both endogenous HO-1 and transiently transfected HO-1 reporter constructs in LMH cells, making this a good system for studying the arsenite-mediated cellular stress response. It was somewhat surprising that LPS and cobalt did not induce the pCHO7.1-Luc reporter construct. The lack of a response to LPS may be explained by a cell-type specific difference, e.g. it is not known whether LMH cells have LPS receptors, which would confer responsiveness to this chemical. The promoter elements required for responsiveness to cadmium and LPS may not be present in this construct, or a silencer element may be preventing reporter gene expression.

The results from this study show that there are clear differences between the activation pathways of arsenite and the other treatments tested (heat shock, cadmium, and heme), since all of these conditions elicited different patterns of MAP kinase activation (Fig. 3). Some researchers have proposed that heme induces HO-1 by a stress-mediated mechanism (4, 7, 13, 34, 35, 84). However, in kinase assays, heme, at concentrations that strongly induce endogenous HO-1, failed to activate any of the MAP kinases (Fig. 3). Therefore, since heme induces endogenous HO-1 in LMH cells (5), it is likely to do so via a non-MAP kinase-mediated pathway. This is consistent with the results of Cable et al. (9), in which heme was proposed to act through a pathway not involving a stress-mediated mechanism.

Two recent studies have investigated the ability of sodium arsenite to activate MAP kinases. One study found that arsenite activated ERK, JNK, and p38 MAP kinases in PC12 cells (59). A separate study in HeLa cells found that JNK and p38 MAP kinases were activated by arsenite; however, no activation of ERK was observed (60). The discrepancy in these results may be explained by the fact that these studies were done in two different cell types (59, 60). It is generally acknowledged that MAP kinases can be differentially regulated by the same stimuli in diverse cell types. Our results in LMH cells indicate that arsenite activates ERK, JNK, and p38 (Figs. 3 and 4), in agreement with the studies done by Liu et al. in Rat-1 and PC-12 cells (59).

Our results suggest that ERK and p38 mediate the MAP kinase induction of HO-1, and that JNK is unlikely to be involved (Figs. 5-9). The activated components of the ERK pathway (Ras, Raf, MEK) and an activated component of the p38 pathway (MKK6) were able to induce luciferase gene expression from the pCHO7.1-Luc reporter construct. In contrast, components upstream of JNK, MEKK1 and MLK3, were not capable of inducing expression of this HO-1 promoter-reporter construct (Figs. 5 and 8). MEKK1 has been reported to activate a number of kinases, including JNK, IB kinase, and ERK in some experimental systems (but only when overexpressed) (69, 85, 86). Our studies show that MEKK1 was unable to induce luciferase gene expression; therefore, these pathways of activation are not likely to be involved in mediating the arsenite induction of HO-1.

Some evidence suggests that gene induction by sodium arsenite is mediated through an AP-1 element (14, 65). Guyton et al. (63) found that arsenite treatment increased nuclear extract binding to an AP-1 element in the GADD153 (CHOP) gene promoter. In our studies of transfected heme oxygenase-1 deletion constructs, increased induction by arsenite correlated with the presence of putative AP-1 consensus sites in the distal promoter. Only a single consensus AP-1 site is located in the first 3.6 kb of the promoter. Three more putative AP-1 sites are found in the promoter region from 3.6 to 7.1 kb. Additionally, there are several other putative promoter elements (TRE, Myc/Max, CREB, and C/EBP sites, as well as an SRE site) in the distal promoter region that could be responsive to MAP kinases. We have shown that at least one AP-1 element, located at 1576 base pairs upstream of the transcription start site, plays a role in the arsenite-mediated induction of HO-1 gene expression (Fig. 10). Given the complexity of the HO-1 promoter, it is likely that other transcription factor elements are also involved in the modulation of HO-1 gene transcription.

Experiments involving dominant negative components of the ERK and p38 pathways (Figs. 5 and 6), and the inhibitors of MEK (PD98059) and p38 (SB203580) (Figs. 7 and 9) provide strong evidence that these two pathways play essential roles in HO-1 gene expression in the presence of arsenite. The combination of MEK and p38 inhibitors was unable to produce a complete block of arsenite induction of pCHO7.1-Luc reporter gene expression. Due to amplification in kinase signaling pathways, a small amount of active kinase may account for the induction observed in the presence of both inhibitors (Fig. 9). However, arsenite signaling is complex and probably also utilizes a pathway that does not involve a phosphorylation event. Our data indicate that most of the arsenite-mediated induction of HO-1 gene expression is transduced by the ERK and p38 MAP kinase pathways (Figs. 5-9). From studies using stably transfected PC12 and transiently transfected Rat-1 cells, Liu et al. (59) suggested that the cellular response to arsenite is partially regulated by a Ras-dependent mechanism and partially by a Ras-independent mechanism. A response involving both the Ras-dependent ERK and the Ras-independent p38 pathways would be consistent with this result.

Several possible mechanisms through which arsenite activates both ERK and p38 would provide an explanation for co-operative activation of HO-1 gene induction: 1) ERK and p38 may target transcription factors that bind to separate promoter elements required for the tightly controlled expression observed for HO-1, 2) ERK and p38 may activate transcription factors that bind to a single promoter element, or 3) ERK and p38 may target a shared downstream kinase. AP-1 elements are a potential target for the combined action of ERK and p38. ERK can increase c-Fos expression in vitro, while p38 activates ATF2, and also increases c-Fos expression. Each kinase activating one of its substrates would lead to formation or activation of AP-1 transcription factor complexes (54, 55, 87-90). There are several putative Myc/Max sites located in the heme oxygenase-1 promoter that may also serve as sites for signal integration by ERK and p38 (65).

Recent studies have established mechanisms that allow integration of the ERK and p38 MAP kinase pathways. Novel protein kinases that are activated by both ERK and p38 MAP kinases have been described (91, 92). These kinases, termed MNK1 and MNK2, are proposed to be a convergence point for co-operative action of the growth factor-regulated ERK pathway and the stress-regulated p38 pathway (91, 92). A second example is provided by Ets transcription factors that are phosphorylated and activated by both ERK and p38 MAP kinases (67, 93). Similar mechanisms may contribute to the integration of the ERK and p38 MAP kinase pathways at the HO-1 promoter.

Another explanation, requiring multiple mechanisms of MAP kinase activation for the full response to arsenite, may involve effects of this chemical on phosphatase activities. Although our evidence supports a role for ERK and p38 in HO-1 expression, these studies do not exclude the possibility that one or both of these MAP kinases may be activated through phosphatase inhibition, rather than by the upstream kinase cascade. Phosphatases have critical sulfhydryl residues that are susceptible to the type of oxidation mediated by arsenite. Inhibition of phosphatases may also lead to an increase in MAP kinase activity. In a study of MAP kinase activation by arsenite, Cavigelli et al. (60) suggested that arsenite may stimulate AP-1 activity by inhibiting a JNK phosphatase. Further investigation into a phosphatase inhibition mechanism for HO-1 induction is currently under way.

In summary, we have shown that arsenite induces both endogenous HO-1 mRNA and transfected HO-1 promoter-driven luciferase gene expression. In addition, arsenite induces the MAP kinases that are involved in cellular responses to mitogens and cellular stressors. An AP-1 element is responsible for part of the arsenite-mediated HO-1 induction. Furthermore, disruption of the ERK or p38 MAP kinase pathways by dominant negatives or inhibitors abrogates arsenite-mediated induction of HO-1. In contrast, JNK does not seem to play a role in this signaling pathway.