HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Seldon, M. P. Articles by Soares, M. P. Search for Related Content PubMed PubMed Citation Articles by Seldon, M. P. Articles by Soares, M. P.

Heme Oxygenase-1 Inhibits the Expression of Adhesion Molecules Associated with Endothelial Cell Activation via Inhibition of NF-B RelA Phosphorylation at Serine 276¹

Mark P. Seldon^2,*, Gabriela Silva^2,*, Nadja Pejanovic^*, Rasmus Larsen^*, Isabel Pombo Gregoire^*, Josina Filipe^*, Josef Anrather and Miguel P. Soares^3,*

^* Instituto Gulbenkian de Ciência, Apartado 14, Oeiras, Portugal; and Department of Neurology and Neuroscience, Division of Neurobiology, Weill Medical College of Cornell University, New York, NY 10021

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
Heme oxygenase-1 (HO-1; encoded by the Hmox1 gene) catalyzes the degradation of free heme into biliverdin, via a reaction that releases iron (Fe) and carbon monoxide. We report that HO-1 down-regulates the proinflammatory phenotype associated with endothelial cell (EC) activation by reducing intracellular nonprotein-bound Fe (labile Fe). EC isolated from Hmox1^–/– mice have higher levels of intracellular labile Fe and reactive oxygen species (ROS) as compared with EC isolated from Hmox1^+/+ mice. Basal and TNF-induced expression of VCAM-1, ICAM-1, and E-selectin were increased in Hmox1^–/– vs Hmox1^+/+ EC, an effect reversed by Fe chelation using deferoxamine mesylate (DFO). Fe chelation inhibits TNF-driven transcription of Vcam-1, Icam-1, and E-selectin, as assessed using luciferase reporter assays. This effect is associated with inhibition of the transcription factor NF-B via a mechanism that is not associated with the inhibition of IB phosphorylation/degradation or NF-B (i.e., RelA) nuclear translocation, although it affects very modestly NF-B binding to DNA B consensus sequences in the Vcam-1 and E-selectin promoters. HO-1 inhibits NF-B (i.e., RelA) phosphorylation at Ser²⁷⁶, a phosphoacceptor that is critical to sustain TNF-driven NF-B activity in EC. This effect was mimicked by Fe chelation as well as by antioxidants (N-acetylcysteine). In conclusion, we demonstrate a novel mechanism via which HO-1 down-modulates the proinflammatory phenotype of activated EC, i.e., the inhibition of RelA phosphorylation at Ser²⁷⁶.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
Under homeostasis, endothelial cells (EC)⁴ regulate blood flow and orchestrate leukocyte trafficking throughout tissues via the expression of genes that regulate vasoconstriction and thrombosis and inhibit leukocyte adhesion (reviewed in Refs. 1 and 2). Under inflammatory conditions, however, EC undergo profound phenotypic modifications, becoming highly vasoconstrictive, prothrombotic, and proadhesive, a phenomenon referred to as EC activation (reviewed in Ref. 1). These phenotypic modifications rely on the down-regulation of vasodilatory, antithrombotic, and antiadhesive genes as well as on the induction of several immediate early responsive vasoconstrictive, prothrombotic, and proadhesive genes (2). Induction of these proinflammatory genes occurs through a mechanism that requires the activation of the transcription factor NF-B (3, 4).

The NF-B family of transcription factors is composed of five members sharing a consensus Rel homology domain (RHD), i.e., p50, p52, cRel, RelB, and RelA (5), the latter of which is probably the predominant active form in EC (6). NF-B proteins form homodimers and heterodimers that can bind B DNA motifs in the promoter regions of "NF-B-dependent genes" (7). In quiescent EC, NF-B dimers are maintained mostly in the cytoplasm by virtue of their interaction with inhibitors of NF-B (IB) molecules that occlude the nuclear localization signals (8, 9) in their RHD. However, when exposed to proinflammatory agonists such as TNF, IBs are phosphorylated, polyubiquitinated, and degraded via the 26S proteosome pathway (10) so that nuclear localization signal exposure on NF-B dimers allows for nuclear translocation and binding to DNA B motifs (reviewed in Ref. 9).

There are additional mechanisms that modulate NF-B transcriptional activity (reviewed in Ref. 11). These include, but are not restricted to, phosphorylation of Ser²⁰⁵ (12), Ser²⁷⁶ (13, 14, 15), Ser²⁸¹ (12), and Ser³¹¹ (16) in the N-terminal domain of RelA. Modulation of these phosphoacceptors is under the control of several kinases, including protein kinase A (PKA) (14), mitogen- and stress-activated kinase-1 (MSK1) (15), and protein kinase C (PKC) (13) (16). The phosphorylation status of RelA controls not only its transcriptional activity but also specificity for different subsets of target genes (12). How these phosphoacceptors modulate RelA activity is not clear but presumably involves the interaction of RelA with coactivators such as the CREB-binding protein (CBP)/p300 (14, 17).

Although essential to elicit inflammatory responses, the expression of proinflammatory genes associated with EC activation must be tightly regulated so as to prevent unfettered inflammation. One of the mechanisms via which this occurs relies on the expression of "protective genes" (18). These have a dual role in that they regulate not only NF-B activity but in addition protect EC from undergoing apoptosis (18). We have previously demonstrated that the stress responsive enzyme heme oxygenase-1 (HO-1, encoded by the Hmox1 gene) acts in such a manner (19, 20).

When exposed to proinflammatory agonists, EC up-regulate the expression of HO-1 via a mechanism that probably requires the relief of its constitutive transcriptional repression by the broad-complex tramtrack and bric a brac (BTB) and cap’and’collar (CNC) homologue 1 (BACH-1) (21, 22). Once BACH-1 is released from the Hmox1 promoter, binding of the transcription activator NF-E2-related factor-2 (NRF-2) occurs, leading to Hmox1 transcription (22).

Under inflammatory conditions, HO-1 becomes the rate-limiting enzyme in the catabolism of heme, a reaction that releases free iron (Fe) from the inner core of the protoporphyrin ring of heme and generates biliverdin as well as the gas carbon monoxide (23). The free Fe generated in this manner up-regulates the expression of heavy chain ferritin (H-ferritin) (24), which forms a multimeric complex with the light chain ferritin and nucleates free Fe, impairing it from partaking in the generation of reactive oxygen species (ROS) via the Fenton reaction (25), a highly deleterious effect (26, 27, 28). Free Fe also induces the expression/activity of an ATPase Fe efflux pump that decreases cellular Fe content (29).

Biliverdin, another end product of heme catabolism by HO-1 (23), is converted into the antioxidant bilirubin (30) by biliverdin reductase (31). Although bilirubin can inhibit NF-B activation in EC (20), the relatively low level of biliverdin reductase expressed in EC makes it unlikely that this antioxidant would act as an autocrine manner to regulate NF-B activity. Therefore, we reasoned that HO-1 might control NF-B activation in EC via its ability to down-modulate the levels of intracellular labile Fe. We provide evidence that the reduction of EC labile Fe content associated with HO-1 expression inhibits specifically the phosphorylation of RelA at Ser²⁷⁶ and probably at Ser²⁰⁵, thus reducing NF-B transcriptional activity. This mechanism underlies the inhibition of proinflammatory gene expression associated with the expression of HO-1 in EC and should contribute in a critical manner to the resolution of inflammatory reactions such as afforded by HO-1 expression.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
Mice

BALB/c Hmox1^–/– mice, generated by S.-F. Yet of the Pulmonary and Critical Care Division, Brigham and Women’s Hospital (32), were maintained at the Instituto Gulbenkian de Ciência (Oeiras, Portugal) under specific pathogen-free conditions in accordance with guidelines from the Animal Use and Institutional Ethical Committee of this institute. Genotypes were verified by genomic PCR using the following primers: Hmox1 (5'-GGTGACAGAAGAGGCTAAG-3' and 5'-CTGTAACTCCACCTCC AAC-3') and neomycin (5'-TCTTGACGAGTTCTTCTGAG-3' and 5'-ACGAAGTGACGCCATCTGT-3').

Cell culture

Murine endothelial cells (MEC) were isolated as described (33). Briefly, three hearts were mechanically minced, digested (2.2U/ml collagenase A at 37°C for 30 min; Sigma-Aldrich), washed, and labeled with rat anti-mouse PECAM-1 Ab (CD31; BD Pharmingen) plus goat anti-rat IgG microbeads MACS (Miltenyi Biotec). Labeled EC were purified in magnetic separation columns as per the manufacturer’s instructions (Miltenyi Biotec). EC phenotype was assessed by cobblestone morphology and confirmed by flow cytometry (CD31/CD105 expression). MEC were passaged 1/3 and used between passages 6 and 12. Bovine aortic endothelial cells (BAEC) were obtained as previously described (19) or purchased from Vec Technologies or Cell Systems and cultured on gelatinized plates in complete medium (MCDB-131; 20% FCS, 20 mM L-glutamine, 50 U/ml penicillin, and 50 µg/ml streptomycin; Invitrogen Life Technologies). BAEC were passaged 1/3 and used between passages 6 and 10. Immortalized human epithelial kidney 293 (HEK-293) cells were obtained from American Type Culture Collection and cultured as for BAEC. Pooled HUVEC (Clonetics) were cultured in EBM-2 medium supplemented according to the provider’s instructions (Clonetics). HUVEC were passaged 1/3 and used between passages 4 and 6. Mouse embryonic fibroblasts (MEF) from RelA^–/– mice were provided by Dr. A. Beg from Columbia University, New York, NY (34). MEF were cultured in DMEM (Mediatech) supplemented with FCS (10%), penicillin (50 U/ml), and streptomycin (50 µg/ml) (Atlanta Biologicals). RelA^–/– MEF were stably transfected with human RelA recombinant retroviruses essentially as described (12).

Immunoprecipitation

Cells were lysed in 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 5 mM MgCl₂, 5 mM KCl, 1% Triton X-100, and a Roche Protease Inhibitor Cocktail (complete EDTA-free, one tablet per 10 ml of lysis solution). Soluble fractions were incubated with a protein G-Sepharose bead/antibody complex in borate-buffered solution with 1% Triton X-100 (for 4 h at 4°C). RelA was immunoprecipitated with anti-mouse mAb (2 µg; catalog no. SC-8008, Santa Cruz Biotechnology). Immunoprecipitated proteins were washed in lysis buffer, resuspended in 2x sample buffer, heated to 100°C for 5 min, resolved by SDS-PAGE, and subjected to Western blotting.

Western blots

Cells were lysed in Laemmli sample buffer and proteins were resolved by electrophoresis on 8–10% SDS-polyacrylamide gels (35), transferred to polyvinylidene fluoride membrane (Bio-Rad), and blotted with Abs to HO-1 (catalog no. SPA-895 or SPA-896, Stressgen), RelA (catalog no. SC-372, Santa Cruz Biotechnology), phospho-Ser536-RelA (catalog no.3031, Cell Signaling), phospho-Ser276-RelA (catalog no.3037, Cell Signaling), IB (catalog no. SC-371, Santa Cruz Biotechnology), phospho-Ser³²-Ser³⁶-IB (catalog no.9246, Cell Signaling), c-Myc (clone 9E10; American Type Culture Collection), or -tubulin (clone DM 1A; catalog no. T-9026, Sigma-Aldrich). Primary Abs were detected using HRP-conjugated secondary Abs (Pierce). Labeling was visualized using SuperSignal West Pico Chemiluminescent Substrate (Pierce) and exposed to BioMax film (Kodak).

Flow cytometry

Confluent EC (4 x 10⁵) were washed in PBS and detached with 0.125% trypsin/0.05% EDTA (at 37°C for 5 min; Invitrogen Life Technologies). Digestion was stopped with complete MCDB-131 medium and cells were pelleted by centrifugation (200 x g at 4°C for 3 min), resuspended in 150 µl of FACS medium (PBS, 3% FCS, and 0.01% sodium azide), transferred into round-bottom 96-well plates (TPP) and pelleted (200 x g at 4°C for 3 min). Supernatant was removed and cells were incubated in 25–50 µl of FACS medium with primary Ab (5 µg/ml for 30 min at 4°C) against endoglin/CD105 (catalog no.550546, BD Biosciences), PECAM/CD31 (catalog no.553370, BD Biosciences), or VCAM-1/CD106 (catalog no.553331, BD Biosciences). Cells were washed in FACS medium (3x 200 µl at 200 x g and 4°C for 3 min) and primary Abs were detected using FITC-labeled secondary Abs (catalog no. F-6258, Sigma-Aldrich) in FACS medium (25 µl for 30 min at 4°C). Cells were costained with propidium iodide (0.5 µg/ml) to assess cellular viability. Fluorescence was measured by flow cytometry on a FACSCalibur device (BD Biosciences).

Cellular Fe assay

Labile and nonlabile Fe were measured essentially as described previously (36) using Ferene-S (37). Briefly, confluent EC (1 x 10⁷) were harvested by trypsinization, pelleted (200 x g at 4°C for 3 min), washed three times with Fe-depleted PBS (1% Chelex-100 for 12 h) (Bio-Rad), and digested (0.5 ml of 25% perchloric acid at 4°C for 30 min). Supernatants (containing labile Fe) were collected after centrifugation (10,000 x g at 4°C for 15 min), pellets were digested (0.1 ml of 10 M nitric acid at 60°C for 16 h), pH was neutralized (10 M NaOH), cellular debris was pelleted (10000 x g at 4°C for 15 min), and supernatants were collected as the nonlabile cellular Fe fraction. Fe in both fractions was converted to the ferrous (Fe²⁺) form using sodium ascorbate (0.25 M, 0.1 ml) plus an equal volume of ammonium acetate (40%). Ferene-S (600 µM; BioVectra) was added and OD was measured at _{594 nm}. Fe levels were normalized to total protein content as determined by Bradford assay. Fe fractions are expressed as moles of Fe per milligram of cellular protein. Assays were performed at least in triplicate for each condition.

Quantitative real-time PCR

cDNA was obtained from total RNA as described (38). The relative number of mRNA transcripts was assessed using LightCycler real-time quantitative PCR (Roche) using FastStart DNA SYBR Green I mix (Roche). The transcript number was calculated using a 2^–CT method (relative number) (39) or quantified as an absolute value using standard cDNA plasmids encoding the same sequences as the ones amplified by PCR. Results were normalized to hypoxanthine-guanine-phosphoribosyltransferase (HPRT) (40). PCR primers used were as follows: 5'-TGGGAGAGACAAAGCAGAAG-3' and 5'-AGATGGTCAAAGGGATACAC-3' or 5'-TGCCGAGCTAAATTACACATTG-3' and 5'-CCTTGTGGAGGGATGTACAGA-3' for Vcam-1; 5'-GCCTTGGTAGAGGTGACTGAG-3' and 5'-GACCGGAGCTGAAAAGTTGTA-3' for Icam-1; 5'-CTGCTCTTGTTTTTGTTCTC-3' and 5'-CTTCCATAGTCAGGGTGTTC-3' for E-selectin; and 5'-GCTGGTGAAAAGGACCTCT-3' and 5'-CACAGGACTAGAACACCTGC-3' for Hprt.

Reporters and expression constructs

Rattus norvegicus Hmox1 cDNA was expressed in pcDNA3 (Invitrogen Life Technologies) (41). N-terminal Myc-tagged RelA, p50, and cRel expression vectors have been described in Refs. 13 and 42 , respectively. The NF-B luciferase (luc) reporter (B-luc) consists of three NF-B binding sites, i.e., NF-E-selectin, B-2, and B-3, derived from the porcine E-selectin promoter (43). Other NF-B luciferase reporter constructs, i.e., VCAM-1.2-B-luc, ICAM-1-B-luc, and E-selectin-B-luc have been described elsewhere (12). Briefly, the minimal SV40 promoter in the pGL3-promoter vector (Promega) was replaced with a minimal 5'-GATCTGGGTATATAATGGATCCCCGGGTACGCAGCTCAAGCT-3' promoter, and the synthetic oligonucleotides coding for the B consensus sequence derived from the Vcam-1, Icam-1, or E-selectin promoter were cloned upstream of this minimal promoter so that they differ only in their decameric B consensus sequences. The prototypical sequence used was 5'-GCT-B-decamer-CTGAGCTCCT-B-deamer-CTCAGCT-3', in which the decameric B consensus were GGGATTTCCC for VCAM-1.2-B-luc, TGGAAATTCC for ICAM-1-B, and GGGGATTTCC for E-selectin-B-luc (12). The human VCAM1 (–1716/+119) luciferase reporter was kindly provided by W. C. Aird (Department of Molecular Medicine, Beth Israel Deaconess Medical Center, Boston, MA) (44). The porcine E-selectin (–1286/+482) luciferase reporter has been described (45). The ICAM-1-luc (–339/-30) luciferase reporter was kindly provided by J. S. Pober (Boyer Center for Molecular Medicine, Yale University School of Medicine, New Haven, Connecticut) (46). The tetracycline operon (tet°) luciferase reporter (tet°-luc) contains seven upstream tet° sequences derived from the tet° flanked by two divergently orientated cytomegalovirus minimal promoters driving the full-length firefly luciferase gene (H. Bujard, University of Heidelberg, Germany) (13). N-terminal Myc-tagged RelA and serine to alanine (S205A, S276A, S281A, and S311A) RelA mutants have been described in Ref. 12 . Myc-tagged RelA chimeric expression vectors have been described elsewhere (13). Briefly, RelA (2–320)/VP16 is composed of the RelA RHD (aa 2–320) fused to virion protein 16 (VP16) (13). TET/RelA (aa 2–551) is composed of a tet° DNA binding domain (TET) fused to the full-length RelA (13). TET/RelA (268–551) is composed of TET fused to the RelA transactivation domain (TAD) (aa 268–551) (13). TET/VP16 is composed of TET fused to VP16 (13). The pSV-β-galactosidase reporter (Promega) consists of the lacZ gene from Escherichia coli under the control of the SV40 early promoter and enhancer, providing for constitutive expression in EC. The pCAGGS-AFP vector drives enhanced GFP expression by a CMV/actin hybrid promoter (47). The pRL-SV40 control vector contains the cDNA encoding Renilla luciferase under the control of the SV40 early promoter and enhancer (Promega).

Recombinant adenoviruses

HUVEC were transduced with recombinant adenoviruses as described (20). The recombinant β-galactosidase adenovirus was a gift from R. Gerard (University of Texas Southwest Medical Center, Dallas, TX). The recombinant HO-1 adenovirus containing the entire coding region of the R. norvegicus HO-1 cDNA has been described (20).

Transient transfection and luciferase assays

BAEC, HUVEC, and HEK-293 cells were transfected at 60–70% confluence using the Lipofectamine 2000 reagent according to the manufacturer’s instructions (Invitrogen Life Science Technologies). In BAEC and HEK-293, 3 µg of total DNA was used per well of a 6-well plate with 2 µl of Lipofectamine 2000. For HUVEC, 1.5–2.5 µg of DNA was used with 1 µl of Lipofectamine 2000. The total amount of DNA was kept constant using pcDNA3 (Promega). Briefly, cells were washed and placed in serum and antibiotic-free medium followed by the addition of the DNA/Lipofectamine and replaced thereafter (2.5 h) with medium containing sera. Transfection efficiency was determined by either flow cytometry or fluorescence microscopy in cells cotransfected with pCAGGS-AFP. Transfection efficiency was typically 50% for BAEC, >95% for HEK-293, and 5–20% for HUVEC.

Luciferase assays were performed using a single (firefly) luciferase assay system (Promega) according to manufacturers instructions with a β-galactosidase control expression vector (pSV-β-galactosidase) as the control for transfection and extraction efficiency. β-Galactosidase expression was detected using Galacto-Light Plus β-Galactosidase Reporter Gene Assay System (Applied Biosystems). Alternatively, a Dual Luciferase Assay System (Promega) was used with a Renilla control expression vector (pRL-SV40). Assays were performed in triplicate and luciferase values were normalized to either β-galactosidase or Renilla luciferase.

Cellular ELISA

Two days postconfluence, HUVEC were exposed to human recombinant TNF at the indicated concentrations and times. The procedure has been described in detail elsewhere (20).

Immunostaings

Confluent EC were washed in PBS and fixed with 3.7% paraformaldehyde (20 min at room temperature). Immunostainings were performed using an anti-RelA mAb (catalog no. SC-372, Santa-Cruz Biotechnology) plus a FITC-labeled goat anti-rabbit polyclonal Ab (Pierce). DNA was stained using 20 ng/ml 4',6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich). Stainings were detected using a fluorescence microscope (Leica DMRA2) at Ex_{480/40 nm} and Em_{527/30 nm} for FITC and Ex_{360/40 nm} and Em_{470/40 nm} (where Ex is excitation and Em is emission) for DAPI. Images were acquired using Metamorph version 4.6r5 software (Universal Imaging Corporation) and analyzed with NIH ImageJ software.

Reagents

Human recombinant TNF (R & D Systems) was resuspended in PBS plus 0.1% BSA and added directly to culture medium. Deferoxamine mesylate (DFO) (Sigma-Aldrich) was dissolved in culture medium at 30 and 250 µM. N-(2-hydroxybenzyl)-L-serine (HBSer; 0.5–5 mM) (48) was a gift from R. Tyrrell, Department of Pharmacology, University of Bath, Bath, U.K. (49). Salicylaldehyde isonicotinoylhydrazone (SIH; 5–100 µM) (gift from Dr. P. Ponka, McGill University, Montreal, Quebec, Canada) was prepared as described (50). N-acetyl-L-cysteine (Sigma-Aldrich) was dissolved in culture medium, neutralized to pH 7.5 with 0.2–1 M NaOH, and added to culture medium (0.5 and 25 mM). 5-(and-6)-Chloromethyl-2',7'-dichlorofluorescein diacetate, acetyl ester (CM-H₂DCFDA, Molecular Probes) was dissolved in DMSO (Sigma-Aldrich) and added to culture medium (3 µM).

ROS detection

Cellular ROS content was assayed by flow cytometry using CM-H₂DCFDA. Two days postconfluence, MEC were collected by trypsinization, pelleted (200 x g at 4°C for 3 min), washed in PBS, and incubated with CM-H₂DCFDA (3 µM in PBS; 37°C for 15 min). Cells were washed twice in PBS with 2% FCS and resuspended in PBS with 2% FCS (37°C for 30 min). Fluorescence was measured by flow cytometry (FACSCalibur; BD Biosciences) and data were analyzed using FlowJo software (Tree Star). Propidium iodide (5 µg/ml) was used to exclude dead cells.

EMSA

Nuclear extracts were obtained from confluent BAEC as described (13, 42). Protein concentration was determined by Bradford assay. EMSA was performed essentially as previously described (13, 42). Briefly, double-stranded oligonucleotides were generated with the following sequences. 5'-AGTTGAGGGGTTTCCCCCAGGC-3' (VCAM-1.1); 5'-AGTTGAGGGGATTTCCCCAGGC-3' (VCAM-1.2); 5'-AGTTGAGTGGAAATTCCCAGGC-3' (ICAM-1); 5'-AGTTGAGGGGGATTTCCCAGGC-3' (E-selectin); and 5'-AGTTGAGGGGACTTTCCCAGGC-3' (Ig-B). Oligonucleotides were labeled by phosphorylation with [-³²P]ATP. EMSA was visualized in the form of autoradiographs indicating DNA/protein binding.

Statistical analysis

Statistical significance was assessed using an unpaired two-tailed t test assuming unequal variances as specified in Results according to experimental design, except for data in Fig. 1e where a one-sample, one-tailed z test was used.

View larger version (28K): [in this window] [in a new window]   FIGURE 1. HO-1 regulates VCAM-1 expression via modulation of EC labile Fe content. a, Endoglin/CD105 was detected by flow cytometry in MEC from Hmox1+/+ or Hmox1–/– BALB/c mice. Gray histograms represent control IgG staining and black histograms represent anti-CD105 staining. b, Hmox1+/+, Hmox1+/–, and Hmox1–/– genotypes were verified by PCR. Neo, Neomycin. c, HO-1 protein was detected by Western blotting using whole cell extracts from Hmox1+/+ and Hmox1–/– MEC. d, Labile and total Fe were quantified in Hmox1+/+ and Hmox1–/– MEC. Results are shown as mean ± SD from six independent samples in one of two independent experiments. e, Vcam-1 mRNA expression in the liver of Hmox1+/+ and Hmox1–/– mice quantified by quantitative real-time PCR. Values are normalized to Hprt mRNA and shown as mean fold increase ± SD relative to Hmox1+/+ (n = 3 pairs of matched animals for each genotype). Statistical significance was assessed using one-sample z test. f, Vcam-1 mRNA expression in EC from Hmox1+/+ or Hmox1–/– mice quantified by quantitative real-time PCR. Values are normalized and shown as in e (n = 3 independent wells in one of two experiments). g, Surface VCAM-1 expression in EC from Hmox1+/+ or Hmox1–/– mice treated or not treated (NT) with DFO (250 µM; 16 h) was detected by flow cytometry. Results shown are geometric mean fluorescence ± SD from triplicate samples in one of three independent experiments. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ns, not significant (p > 0.05).

	Results

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

HO-1 deficient mice (51) and humans (52) have deregulated Fe metabolism and widespread EC activation/injury, suggesting that HO-1 expression is essential in the regulation of Fe homeostasis (Ref. 51 and reviewed in Ref. 53) and in preventing unfettered EC activation. Because a causal link between deregulated Fe metabolism and unfettered EC activation has not been established, we asked whether deregulated Fe metabolism would promote EC activation. Primary EC were isolated from Hmox1^+/+ and Hmox1^–/– mice and their phenotype was confirmed by flow cytometry, i.e., expression of endoglin/CD105 (Fig. 1a). Genotype was confirmed by PCR (Fig. 1b) and Western blotting (Fig. 1c). Labile Fe content increased by 23% in Hmox1^–/– vs Hmox1^+/+ EC (Fig. 1d), while nonlabile Fe content was similar whether or not HO-1 was expressed (Fig. 1d), suggesting that HO-1 regulates specifically the labile Fe content of EC.

We then asked whether increased labile Fe content resulting from Hmox1 deletion would modulate the expression of adhesion molecules in EC. Expression of Vcam-1 mRNA, an adhesion molecule associated with EC activation (reviewed in Refs. 1 and 2), was increased by 3- to 4-fold in the liver of Hmox1^–/– vs Hmox1^+/+ mice as quantified by quantitative real-time PCR (Fig. 1e). To ascertain that HO-1 expression in EC accounted for this effect, Vcam-1 mRNA expression was quantified specifically in EC isolated from Hmox1^–/– or Hmox1^+/+ mice. The basal level of Vcam-1 mRNA expression increased by 4- to 5-fold in EC from Hmox1^–/– vs Hmox1^+/+ mice (Fig. 1f). Expression of VCAM-1 protein was also significantly increased in Hmox1^–/– vs Hmox1^+/+ EC (Fig. 1g). This effect was probably due to higher levels of labile Fe in Hmox1^–/– EC (Fig. 1d), as Fe chelation by DFO reduced VCAM-1 expression in Hmox1^–/– EC to levels similar to those of Hmox1^+/+ EC (Fig. 1g). Taken together, these observations suggest that HO-1 expression in EC might be a physiologic regulator of basal VCAM-1 expression, an effect that probably relies on the control of labile Fe.

Down-regulation of labile Fe associated with the expression of HO-1 inhibits TNF-driven up-regulation of adhesion molecules in EC

We have previously shown that when overexpressed in EC, HO-1 inhibits the expression of adhesion molecules associated with EC activation, including VCAM-1, E-selectin, and ICAM-1 (20). We now asked to what extent this effect is associated with decreased labile Fe content (Fig. 1d). Transient HO-1 overexpression in EC reduced labile Fe content by 17% as compared with control EC (Fig. 2a). A similar effect was observed when nontransfected EC were treated with the Fe chelator DFO, i.e., an 18% reduction in labile Fe content as compared with untreated EC (Fig. 2a). TNF, a well-established proinflammatory agonist, induced a transient increase of 12–20% in EC labile Fe content (Fig. 2b), an effect in keeping with that reported in other cell types (54). HO-1 overexpression (Fig. 2c) or Fe chelation by DFO (Fig. 2d) reduced by 30–40% the ability of TNF to up-regulate VCAM-1 expression in EC. The inhibitory effect of Fe chelation was dose-dependent in that higher DFO concentrations enhanced its effect (Fig. 2e). To exclude the possibility that DFO would function independently of its Fe-chelating activity, experiments were performed using HBSer, a "milder" Fe chelator (49). HBSer mimicked the effect of DFO, but only when TNF concentrations were <1 ng/ml, i.e., a 45% reduction in VCAM-1 expression, as compared with control EC (Fig. 2f). Similar results were obtained using SIH, a cell-permeable Fe chelator (50) (data not shown). These data suggest that despite their relative differences in efficiency, it is the Fe chelation activity of these compounds, i.e., DFO, HbSer, and SIH, that affords their effects.

View larger version (42K): [in this window] [in a new window]   FIGURE 2. HO-1 inhibits TNF-driven expression of VCAM-1, E-selectin, and ICAM-1 via reduction of labile Fe content. a, Labile Fe was quantified in BAEC transiently transfected with control or HO-1 expression vectors as well as in nontransfected BAEC treated or not treated (NT) with DFO (250 µM; 16 h). Mean values ± SD (nine independent samples). b, Labile Fe was quantified in confluent BAEC not-treated (0) or treated with TNF (50 ng/ml). Mean values ± SD from triplicate samples (one of three representative experiments) are shown. c, Nontransduced (NT), β-galactosidase (β-gal), and HO-1 recombinant adenovirus-transduced HUVEC (24 h) were stimulated with TNF (6 h). VCAM-1 was detected by cellular ELISA. d, HUVEC treated with DFO (250 µM; 16 h) and stimulated with increasing concentrations of TNF (6 h). VCAM-1 was detected by cellular ELISA. e, HUVEC were treated with increasing concentrations of DFO (16 h) and stimulated with TNF (10 ng/ml). VCAM-1 expression was assessed as in c and d. Results shown in c–e are mean values ± SD from triplicate samples in one of three independent experiments. f, HUVEC were not treated (NT) or treated with the Fe chelator HBSer (5 mM; 16 h) and not further treated (Cont) or exposed to TNF (0.1 ng/ml; 8 h). VCAM-1 expression was detected as in c–e. Mean values ± SD (triplicate samples in one of four independent experiments). g–i, Confluent MEC from Hmox1–/– or Hmox1+/+ mice were treated with DFO (250 µM; 16 h) and TNF- (50 ng/ml; 3 h) when indicated (+). Vcam-1 (g), E-selectin (h), and Icam-1 (i) mRNA were quantified by quantitative real-time PCR. Values are normalized to Hprt mRNA and shown as the mean number of molecules ± SD (n = 3 independent wells in one representative experiment). *, p < 0.05; **, p < 0.01; ***, p < 0.001; ns, not significant (p > 0.05).

We then asked whether, when expressed under physiological conditions, HO-1 would control the expression of other TNF-responsive adhesion molecules in EC. Up-regulation of E-selectin (Fig. 2h) and Icam-1 (Fig. 2i) mRNA in response to TNF was increased by 3- and 11-fold in Hmox1^–/– vs Hmox1^+/+ EC, respectively. Higher E-selectin (Fig. 2h) and Icam-1 (Fig. 2i) mRNA expression in Hmox1^–/– EC was ablated by DFO, suggesting again that it is the higher labile Fe of Hmox1^–/– EC that accounts for this effect. Taken together, these data suggest that, when expressed under physiological conditions, HO-1 reduces the expression of adhesion molecules associated with EC activation, most probably via down-modulation of labile Fe.

Fe chelation inhibits the transcription of genes encoding adhesion molecules in EC

We tested whether down-modulation of labile Fe inhibits Vcam-1, E-selectin, or Icam-1 transcription. EC were transiently transfected with a VCAM-1-luc reporter and treated or not treated with DFO. Induction of VCAM-1 reporter activity by TNF was reduced by 77–100% in EC treated with DFO vs control EC (Fig. 3a). This effect was dose dependent in that increasing DFO concentrations (60–250 µM) enhanced its inhibitory effect (Fig. 3a). The inhibitory effect of DFO was observed across a wide range of TNF concentrations (0.16–100 ng/ml) (Fig. 3b). Finally, this effect was not associated with detectable cytotoxicity, as assessed by flow cytometry (data not shown).

View larger version (30K): [in this window] [in a new window]   FIGURE 3. Fe chelation inhibits TNF-driven Vcam-1, E-selectin, and Icam-1 transcription. a and b, BAEC were transiently transfected with a VCAM-1 luciferase reporter (46 ) and treated with increasing concentrations of DFO (16 h) and TNF (10 ng/ml; 6 h) (a) or DFO (250 µM, 16 h) and increasing concentrations of TNF (6 h) (b). c and d, BAEC were transiently transfected with an E-selectin luciferase reporter (46 ) and treated as in a (c) or treated as in b (d). e and f, BAEC were transiently transfected with an ICAM-1 luciferase (46 ) reporter and treated as in a (e) or treated as in b (f). Results shown are mean luciferase normalized to β-galactosidase units ± SD from triplicate samples in one of three independent experiments. (*, p < 0.05). NT, Not treated.

Reduction of EC labile Fe content inhibits NF-B activity

Induction of Vcam-1, E-selectin, and Icam-1 transcription by TNF is strictly dependent on activation of the transcription factor NF-B in EC (3, 4, 19, 55). Therefore, we reasoned that the reduction of labile Fe might inhibit NF-B activation (Fig. 3). EC were transiently transfected with a synthetic NF-B-luc reporter derived from the porcine E-selectin promoter (43) plus or minus an HO-1 expression vector. TNF-driven NF-B reporter activity was reduced by 42–57% in EC transfected with HO-1 as compared with control EC (Fig. 4a). The inhibitory effect of HO-1 was dose dependent; i.e., higher HO-1 expression resulted in lower NF-B reporter activity (Fig. 4a). This effect was observed over a broad range of TNF concentrations (1–10 ng/ml) (Fig. 4b). Given that HO-1 overexpression reduces the labile Fe content of EC (Fig. 2a), we asked whether Fe chelation would mimic the inhibitory effect of HO-1. DFO inhibited TNF-driven NF-B reporter activity by 63–78% as compared with control EC not treated with DFO (Fig. 4c). This effect was dose dependent; i.e., higher DFO concentrations resulting in lower NF-B reporter activity (Fig. 4c), and this was observed over a broad range of TNF concentrations (1–10 ng/ml) (Fig. 4d). These data suggest that HO-1 inhibits NF-B activation via a mechanism probably involving the reduction of labile Fe content in EC.

View larger version (35K): [in this window] [in a new window]   FIGURE 4. HO-1 overexpression and Fe chelation inhibit NF-B activation in EC. a, BAEC were transiently cotransfected with a NF-B luciferase reporter derived from the porcine E-selectin promoter (43 ) (B-luc) plus pcDNA3-HO-1 (3000 ng of DNA/3 x 105 cells) and stimulated with TNF (10 ng/ml; 6 h) or not treated (NT). b, BAEC were transiently transfected as in a and stimulated with increasing concentrations of TNF. c, BAEC were transiently transfected with a NF-B luciferase reporter (B-luc), treated with DFO (16 h), and stimulated with TNF (10 ng/ml, 6 h) or not treated (NT). d, BAEC were transiently transfected as in c, treated with DFO (250 µM; 16 h), and stimulated with TNF (6 h). Results shown are mean luciferase normalized to β-galactosidase units ± SD from triplicate samples in one of three independent experiments. *, p < 0.05.

Reduction of labile Fe in EC does not affect IB phosphorylation/degradation or RelA nuclear translocation

When exposed to DFO under the same experimental conditions shown to inhibit NF-B activation (Fig. 4, c and d), TNF-driven IB phosphorylation and/or IB degradation was not affected as compared with control EC (Fig. 5a). DFO also did not alter de novo IB synthesis after TNF stimulation as compared with control EC (Fig. 5a). In addition, DFO failed to modulate the levels of RelA protein expression or RelA phosphorylation at the C-terminal region, i.e., Ser536, as compared with control EC (Fig. 5a). RelA nuclear translocation in response to TNF was also not affected by DFO as compared with control EC (Fig. 5b). Total nuclear RelA was also not significantly modulated by DFO as assessed by Western blot analysis (Fig. 5c) (56).

View larger version (49K): [in this window] [in a new window]   FIGURE 5. Fe chelation does not affect IB phosphorylation/degradation or NF-B nuclear translocation while inhibiting moderately NF-B binding to specific B DNA binding motifs in the Vcam-1, E-selectin, and Icam-1 promoters. a, Confluent BAEC either not treated (Control) or pretreated with DFO (250 µM; 16 h) were stimulated with TNF (10 ng/ml). IB, phosphorylated IB (p-IB), RelA, and Ser536 phosphorylated RelA (p-RelA-Ser536) were detected by Western blotting. Immunoblots are representative of three independent experiments. b, BAEC treated as in a, RelA (green), and DNA (blue) were immunostained with anti-RelA Ab and DAPI, respectively. Images (x400 original magnification) were acquired by fluorescence microscopy and are representative of four fields from one of three representative experiments. c, BAEC were treated as in a. RelA was monitored in cytoplasm and nucleus by Western blotting. d, BAEC were treated as in a. EMSA were performed using nuclear extracts (N.Ext) plus radiolabeled oligonucleotides corresponding to the B binding sites from the human Ig locus (Ig-B), Vcam1 (–167 bp for VCAM-1.1 and –152bp for VCAM-1.2), E-selectin (–99 bp), and Icam1 (–1390 bp) promoters. Specific (I–IV; red arrow) and nonspecific (ns; black arrows) complexes are shown. Binding specificity was assessed using for each probe using 20-fold molar excess of cold probe (*).

Fe chelation interferes modestly with NF-B binding to specific B DNA binding motifs in the promoter region of Vcam-1, E-selectin, or Icam-1

Given that Fe chelation failed to modulate IB phosphorylation/degradation (Fig. 5a) and/or RelA nuclear translocation (Fig. 5, b and c), we reasoned that inhibition of NF-B activity (Fig. 4) might occur "downstream" of nuclear translocation. Binding of NF-B to B DNA binding motifs derived from the Vcam-1, E-selectin, and Icam-1 promoters was compared by EMSA, using nuclear extracts from EC pre-exposed or not pre-exposed to DFO (16 h) and stimulated or not stimulated thereafter with TNF (60 min). DFO failed to modulate NF-B binding to a "standard" B binding motif derived from the Ig promoter (Ig-B) as compared with control EC (Fig. 5d). In a similar manner, DFO failed to modulate the very modest binding of NF-B to the –167-bp Vcam-1 B binding motif (VCAM1.1). There was a slight but reproducible inhibition of NF-B binding to the –152-bp Vcam-1 B binding motif (VCAM1.2) in EC treated with DFO, as compared with control EC (Fig. 5d). DFO reduced very slightly NF-B binding to the –99bp E-selectin B binding motif while not affecting NF-B binding to the –1390-bp Icam-1 B DNA binding motif (Fig. 5d) as compared with control EC. In all cases, the characteristic pattern of NF-B/DNA binding was observed, namely the appearance of "specific bands" upon TNF stimulation that were competed out using 20-fold excess unlabeled B probe, corresponding to the probe used in each EMSA, respectively (Fig. 5d).

Fe chelation inhibits RelA transcriptional activity

Because the inhibition of NF-B activation by HO-1 or Fe chelation (Fig. 4) was not associated with the inhibition of NF-B nuclear translocation (Fig. 5, b and c), and only to a modest extent with DNA binding (Fig. 5d), we tested whether the inhibitory effect of HO-1 acted directly on RelA, presumably the main NF-B transactivator in EC (6). EC from Hmox1^+/+ and Hmox1^–/– mice were transiently transfected with increasing amounts of RelA plus an NF-B luciferase reporter derived from the porcine E-selectin promoter (43). RelA activity was 2–3-fold higher in EC from Hmox1^–/– mice as compared with EC from Hmox1^+/+ mice (Fig. 6a). This suggests that, when expressed under physiologic conditions, HO-1 inhibits RelA activity.

View larger version (33K): [in this window] [in a new window]   FIGURE 6. Fe chelation inhibits RelA transcriptional activity. a, Hmox1+/+ and Hmox1–/– EC transiently cotransfected with RelA and a NF-B luciferase reporter derived from the porcine E-selectin promoter B-luc (43 ). Mean fold induction of luciferase/β-galactosidase (compared with EC not transfected with RelA) ± SD (triplicates in one of two independent experiments). b, BAEC transiently cotransfected with the same NF-B luciferase reporter as in a plus control pcDNA3 (Cont), RelA, RelA plus p50, or RelA plus cRel. Cells were either not treated (NT) or treated with DFO (250 µM; 16 h). Mean fold luciferase/β-galactosidase induction vs control (Cont) EC ± SD (triplicate samples in one of three independent experiments). c, Overexpressed RelA was detected by Western blotting using an Ab recognizing the c-Myc tag in the N terminus of RelA. d, HEK-293 cells were transiently cotransfected and treated as in b. e, RelA–/– MEF were transduced with a control or a RelA recombinant retrovirus and treated or not treated (NT) as in a. mRNA was quantified by quantitative real-time PCR. Mean fold luciferase/β-galactosidase induction vs control (Cont) EC ± SD (triplicate samples in one experiment). f, BAEC were transiently cotransfected and treated with DFO (250 µM; 16 h) or SIH (100 µM; 16 h). Mean fold luciferase/β-galactosidase induction vs control untreated (NT) EC ± SD (triplicate samples in one of three independent experiments). g, Hmox1+/+ or Hmox1–/– EC were pretreated or not treated (NT) with DFO (250 µM; 12–16 h) and exposed or not exposed to TNF (50 ng/ml; 1 h). ROS were detected by flow cytometry using CM-H2DCFDA. Geo means ± SD (one triplicate sample in one of four independent experiments). h, BAEC were transiently transfected as in f and, when indicated, treated with NAC (25 mM; 16 h). Mean fold luciferase/β-galactosidase induction vs control untreated (NT) EC ± SD (triplicate samples in one of three independent experiments). *, p < 0.05; **, p < 0.01; ***, p < 0.001.

We then asked whether Fe chelation would inhibit endogenous Vcam-1, E-selectin, and Icam-1 transcription when driven specifically by RelA. For this purpose, RelA^–/– MEF were transduced or not transduced with a recombinant RelA retrovirus, allowing for the assessment of RelA-driven Vcam-1, E-selectin, and Icam-1 mRNA transcription by quantitative real-time PCR. RelA^–/– MEF reconstituted with RelA expressed Vcam-1, E-selectin, and Icam-1 mRNA transcripts that were 150-, 2500-, and 30.000-fold more abundant to those detected in RelA^–/– MEF, respectively (Fig. 6e). DFO inhibited by 69, 58, and 70% Vcam-1, E-selectin, and Icam-1 mRNA expression driven by RelA, respectively (Fig. 6e). This observation provides direct evidence that Fe chelation inhibits RelA-mediated transcription of endogenous Vcam-1, E-selectin, and Icam-1.

As for DFO, SIH, a cell-permeable Fe chelator (50), also inhibited RelA activity as assessed in EC transiently cotransfected with RelA plus NF-B luciferase reporter (Fig. 6f). This suggests that the inhibitory effect of these molecules is due to their Fe-chelating activity. We reasoned that the reduction of EC labile Fe content might down-regulate RelA activity by disabling labile Fe from promoting the generation of ROS through the Fenton reaction. If this is the case, then HO-1 expression should inhibit ROS accumulation in EC. Intracellular ROS content was 2-fold higher in EC isolated from Hmox1^–/– vs Hmox1^+/+ mice (Fig. 6g), suggesting that when expressed under physiologic conditions, HO-1 limits the accumulation of ROS in EC. Fe chelation by DFO reduced intracellular ROS in Hmox1^–/– EC to levels similar to those of Hmox1^+/+ EC (Fig. 6g). This suggests that the higher levels of intracellular ROS in Hmox1^–/– EC is due to the higher levels of intracellular labile Fe in these cells (Fig. 1d). Similar results were observed using another Fe chelator, i.e., SIH (data not shown). TNF failed to increase ROS levels in EC as compared with untreated EC (Fig. 6g).

If HO-1 inhibits RelA activity by disabling labile Fe from promoting the generation of ROS, then an antioxidant such as N-acetylcysteine (NAC) should mimic the inhibitory effects of HO-1 and/or that of Fe chelation. In keeping with this notion, NAC inhibited by 50% RelA activity as assessed in EC transiently cotransfected with RelA plus NF-B luciferase reporter (Fig. 6h).

Fe chelation targets the N-terminal domain of RelA

We used a series of chimeric DNA constructs encoding the N- or C-terminal domains from RelA that allowed us to test whether Fe chelation inhibits RelA activity via a mechanism that targets its N- and/or C-terminal domains (13). The chimeric constructs were transiently cotransfected in EC with a NF-B luciferase reporter derived from the porcine E-selectin promoter (43). DFO inhibited by 83% the transcriptional activity of a chimeric construct in which the N-terminal domain of RelA was fused to a VP16 TAD (RelA (2–320)/VP16), as compared with control EC not treated with DFO (Fig. 7a). This inhibitory effect was similar to the one exerted over the full length RelA (Fig. 7a). Fe chelation did not modulate the expression of RelA (2–320)/VP16 as assessed by Western blot analysis (Fig 7b). Because RelA (2–320)/VP16 can bind DNA B sequences via the N-terminal domain of RelA and promote gene transcription via the VP16 TAD, these data suggest that Fe chelation targets the N-terminal domain of RelA.

View larger version (21K): [in this window] [in a new window]   FIGURE 7. Fe chelation modulates RelA transcriptional activity via a mechanism that targets the N-terminal domain of RelA. a, BAEC were transiently transfected with an NF-B luciferase reporter derived from the porcine E-selectin promoter (B-luc) (43 ) plus an expression vector encoding the full-length RelA or the N-terminal domain of RelA (aa 2–320) fused to the VP16 TAD (RelA (2–320)/VP16). EC were not further treated (NT) or treated with DFO (250 µM; 16 h). Shown is mean luciferase normalized to β-galactosidase units ± SD (triplicate sample in one of three independent experiments). b, BAEC were transiently cotransfected and treated as in a and overexpressed RelA and RelA (2–320)/VP16 were detected by Western blotting using an anti-c-Myc Ab recognizing the c-Myc tag in the N terminus of these constructs. c, BAEC were transiently transfected with tet°-luc reporter alone (control) or cotransfected with tet°-luc plus TET/VP16 or a construct encoding a TET DNA binding domain fused to the RelA TAD (TET/RelA (268–551). EC were treated as in a. d, BAEC were transiently transfected with NF-B luciferase reporter (B-luc) alone (Control) or cotransfected with B-luc (same as in a) plus an expression vector encoding the TET DNA binding domain fused to the full-length RelA (TET/RelA). EC were treated as in a and b. e, BAEC were transiently transfected with tet°-luc reporter (tet°-luc) alone (control) or cotransfected with tet°-luc plus TET/RelA and treated as in a–c. Data illustrated in a–e are mean luciferase normalized to β-galactosidase units ± SD (triplicate sample in one of three independent experiments). f, BAEC were transiently transfected and treated as in e and TET/RelA was detected by Western blotting using an anti-RelA Ab. *, p < 0.05.

We then asked whether the inhibition of RelA transcriptional activity by Fe chelation targets DNA binding via the RHD. To this aim, we made use of a chimeric DNA construct in which TET was fused via its C-terminal domain to RelA. This chimeric protein (TET/RelA) can transcribe a NF-B luciferase reporter via DNA binding through the RelA RHD (Fig. 7d). The same chimeric protein (TET/RelA) can also transcribe a tet° luciferase reporter via DNA binding through the TET domain (Fig. 7e). Fe chelation by DFO inhibited by 30% TET/RelA-driven NF-B reporter activity (Fig. 7d) but failed to inhibit TET/RelA-driven tet° reporter activity (Fig. 7e). DFO had no effect on TET/RelA expression as assessed by Western blot analysis (Fig. 7f). Taken together, these data suggest that Fe chelation inhibits RelA activity via a mechanism that involves the recognition of B DNA motifs by the RHD.

Fe chelation targets Ser²⁷⁶ and Ser²⁰⁵ in the N-terminal domain of RelA

We hypothesized that Fe chelation might target one or several phosphoacceptors in the N-terminal domain of RelA, an effect that would be consistent with the observed inhibition of RelA activity (reviewed in Ref. 11) (12, 13). We tested this hypothesis specifically for the phosphoacceptors Ser²⁰⁵, Ser²⁷⁶, Ser²⁸¹, and Ser³¹¹. EC were transiently cotransfected with wild-type RelA or with RelA S205A, S276A, S281A, and S311A point mutants plus an NF-B luciferase reporter derived from the porcine E-selectin promoter (43). DFO inhibited by 53 and 43% wild-type RelA and S311A RelA-driven transactivation, respectively, as compared with control EC not treated with DFO (Fig. 8a). This suggests that Fe chelation does not act via Ser³¹¹ to suppress RelA activity. Because the transcriptional activity of the RelA S205A, S276A, and S281A point mutants is abolished using this NF-B reporter (Fig. 8a) (12), we made use of synthetic NF-B reporters consisting of two tandem decameric B sites derived from the human Vcam-1 (VCAM-1.2-B-luc), E-selectin (E-selectin-B-luc), or Icam-1 (ICAM-1-B-luc) promoters (12).

View larger version (50K): [in this window] [in a new window]   FIGURE 8. Fe chelation modulates RelA transcriptional activity via Ser205 and Ser276. BAEC were transiently cotransfected with RelA or RelA S205A, S276A, S281A, or S311A point mutants plus an NF-B luciferase reporter derived from the porcine E-selectin (B-luc) (43 ) (a), human VCAM-1 (VCAM-1.2-B-luc) (12 ) (b), human ICAM-1 (ICAM-1-B-luc) (12 ) (c), or human E-selectin (E-selectin-B-luc) (12 ) (d) promoters. EC were either not further treated (NT) or treated with DFO (250 µM; 24–48 h). The data illustrated in a–d represent mean luciferase normalized to β-galactosidase units ± SD (triplicate in one of at least two independent experiments for each promoter used). e and f, Confluent MEF (e) or MEC (f) were either left untreated (–) or treated (+) with DFO (250 µM; 16 h) and not further treated (–) or treated (+) with TNF (50 ng/ml; 15 min). RelA was immunoprecipitated (IP) and total RelA and RelA Ser276 phosphorylation were detected by Western blotting (WB). Immunoprecipitation with irrelevant Ig was performed as a negative control. *, p < 0.05.

We then asked whether Fe chelation modulates RelA phosphorylation at Ser²⁷⁶, an effect that would justify its inhibitory effect (Fig. 8, a–d) (12, 13). As described previously (13), RelA Ser²⁷⁶ is constitutively phosphorylated in resting MEF (Fig. 8e) as well as in MEC (Fig. 8f). Exposure to TNF failed to increase RelA Ser²⁷⁶ phosphorylation in MEF (Fig. 8e) but did so in MEC (Fig. 8f). Fe chelation by DFO decreased Ser²⁷⁶ phosphorylation by 50–60% in resting as well as in TNF-stimulated MEF (Fig. 8e). Fe chelation fully suppressed TNF-driven Ser²⁷⁶ phosphorylation in MEC (Fig. 8f) while not affecting basal RelA Ser²⁷⁶ phosphorylation (Fig. 8f). Taken together, these observations suggest that down-modulation of labile Fe inhibits RelA Ser²⁷⁶ phosphorylation, thus explaining its ability to suppress RelA activity (12, 13).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
Given that HO-1 overexpression reduces cellular Fe content (29) we asked whether HO-1 would act in a similar manner when expressed under physiologic conditions in EC. Our data suggest that this is the case, because labile Fe content was significantly higher in EC isolated from Hmox1^–/– mice as compared with those isolated from Hmox1^+/+ mice (Fig. 1d). Based on the above, we suggest that HO-1 might act as a physiologic regulator of EC labile Fe content, a notion in keeping with the deposition of labile Fe in the vasculature of Hmox1-deficient mice (51) as well as that associated with the one reported case of HMOX1 deficiency in humans (52).

Several mechanisms have been proposed as to how HO-1 modulates cellular labile Fe content, including through the up-regulation of H-ferritin (24) and/or the activation of ATPase Fe efflux pumps (29). Given that free Fe can induce the expression of H-ferritin, one would predict that increased labile Fe content in Hmox1^–/– EC would lead to higher H-ferritin expression (24). However, H-ferritin expression was very significantly decreased in Hmox1^–/– vs Hmox1^+/+ EC (data not shown). One possible explanation for this apparently contradictory observation would be that the accumulation of ROS in Hmox1^–/– (Fig. 6g) might promote H-ferritin degradation (57), an effect that could account for an overall decrease in H-ferritin expression and thus for the increased labile Fe content observed in Hmox1^–/– EC (Fig. 1d).

EC from Hmox1^–/– mice expressed higher basal levels of VCAM-1 as compared with EC from Hmox1^+/+ mice (Fig. 1, e–g). This suggests that HO-1 is a physiologic regulator of VCAM-1 expression in EC. Increased VCAM-1 expression in Hmox1^–/– EC was reversed by Fe chelation (Fig. 1g), suggesting that there is a functional link between the ability of HO-1 to down-regulate labile Fe and to inhibit VCAM-1 expression.

HO-1 overexpression in EC decreases labile Fe content to the same extent as Fe chelation by DFO (Fig. 2a), an effect that probably explains the ability of HO-1 to inhibit the up-regulation of VCAM-1 expression in response to proinflammatory agonists such as TNF. That labile Fe may be involved in the regulation of proinflammatory gene expression in response to TNF is suggested by the observation that TNF triggers a transient rise of labile Fe content in EC (Fig. 2b). When labile Fe is decreased, either by overexpressing HO-1 or by chelating Fe, the ability of TNF to induce the expression of VCAM-1 is inhibited (Fig. 2, c–f), a finding in keeping with our previous results (20). The physiologic relevance of these findings is supported by the observation that Hmox1 deletion increases TNF-driven VCAM-1 expression, an effect reversed when labile Fe is chelated by DFO (Fig. 2g). Similar effects were observed for other adhesion molecules such as E-selectin (Fig. 2h) and ICAM-1 (Fig. 2i), suggesting that the inhibitory effects of HO-1/Fe chelation can be extended to other proinflammatory genes associated with EC activation.

The observation that Fe chelation inhibits TNF-driven Vcam-1 (Fig. 3, a and b), E-selectin (Fig. 3, c and d), and Icam-1 (Fig. 3, e and f) transcription suggests that NF-B, a transcription factor required for TNF-driven transcription of these genes in EC (3, 20, 55), might be the target of Fe chelation (20). Although our data suggest that this is the case (Fig. 4, a–d), we found that inhibition of NF-B does not occur at the level of IB phosphorylation/degradation (Fig. 5a), C-terminal RelA phosphorylation, i.e., Ser⁵³⁶ (Fig. 5a), or RelA nuclear translocation (Fig. 5b), and most probably not via the inhibition of NF-B binding to B DNA binding motifs in the promoter of Vcam-1, E-selectin, or Icam-1 (Fig. 5d). Fe chelation inhibits RelA transcriptional activity via a mechanism that targets specifically its N-terminal domain (Fig. 7, a and b).

Because N-terminal phosphorylation is required to support RelA transactivation activity (12, 13), we tested whether Fe chelation inhibited RelA activity via a mechanism dependent on serine phosphoacceptors in the N-terminal domain of RelA. Fe chelation failed to inhibit the transcriptional activity of RelA S205A or S276A point mutants as assessed by using synthetic NF-B luciferase reporters containing B DNA motifs from the human Vcam-1 (Fig. 8b), Icam-1 (Fig. 8c), or E-selectin (Fig. 8d) promoters. This led to the hypothesis that HO-1 and/or Fe chelation might interfere with Ser²⁷⁶ phosphorylation. We found that Fe chelation inhibits RelA Ser²⁷⁶ phosphorylation (Fig. 8, e and f) in a manner that should account for its inhibitory effect over RelA activity (12, 13). Whether Fe chelation modulates RelA Ser²⁰⁵ phosphorylation could not be established, as reagents are not available to monitor this phosphoacceptor.

A likely scenario to explain the inhibitory effects observed would be that labile Fe availability, which promotes the generation of ROS (Fig. 6g) via the Fenton reaction (reviewed in Ref. 53), might regulate the expression and/or activity of kinase(s) and/or phosphatase(s) that target RelA at Ser²⁷⁶ and/or Ser²⁰⁵ (Fig. 8, a–d) and thus regulate RelA activity (17). The notion that labile Fe targets RelA activity via ROS is supported by the observation that the antioxidant NAC, a radical scavenger and glutathione precursor, suppressed RelA activity in a manner that mimicked Fe chelation (Fig. 6h). This inhibitory effect is in keeping with our previous observation that bilirubin, another potent antioxidant (30), also inhibits NF-B activation in EC (20). Whether bilirubin targets RelA in a similar manner to Fe chelation, i.e., inhibiting RelA Ser²⁷⁶ phosphorylation, remains to be established but is likely to be the case.

Putative candidate kinases modulated by ROS and targeting RelA activity in EC include PKA (58, 59), mitogen- and stress-activated kinase-1 (MSK1) (15), and protein kinase C (PKC) (12, 13). Among these, PKA, which has been shown to target RelA Ser²⁷⁶, is probably the most likely candidate because Fe chelation can regulate its expression/activity (60). Whether the ability of HO-1 and/or Fe chelation to inhibit RelA activity acts via PKA remains to be established.

We suggest, that the putative kinase(s) and/or phosphatase(s) that target RelA Ser²⁷⁶ interfere with the mechanism via which the RHD promotes transcriptional activity upon binding to B DNA motifs. This is supported by the observation that Fe chelation inhibits the transcriptional activity of a RelA chimeric protein (TET/RelA) only when this protein binds DNA via the RHD (i.e., binding to B DNA recognition motifs in a NF-B reporter; Fig. 7d) but not when the same protein binds DNA via its TET domain (i.e., binding to tet° recognition motifs in a Tet reporter) (Fig. 7e). It is possible that such kinase(s) and/or phosphatase(s) might target RHD after binding to B DNA recognition motifs. This hypothesis remains, however, to be tested.

The mechanism via which HO-1 controls NF-B activation in EC provides clues as to its protective effect in the context of vascular inflammation (61, 62, 63). We suggest that down-regulation of EC labile Fe content may inhibit specifically the expression of proinflammatory genes associated with EC activation, e.g., adhesion molecules, while sparing that of anti-apoptotic genes (42). Such an effect is consistent with targeting of RelA Ser²⁷⁶, a phosphoacceptor that controls RelA cis-acting element specificity (12). That is, by down modulating Ser²⁷⁶ RelA phosphorylation, HO-1 and/or Fe chelation might direct this NF-B family member to transcribe specific subsets of genes (12). It is likely that this would occur via modulation of RelA interaction with coactivators required to support the transcription of specific gene subsets (17). Such a mechanism would explain how HO-1 has reached a "functional compromise" in which it inhibits specifically the transactivation of proinflammatory genes but probably not that of anti-apoptotic genes required to support its cytoprotective effects, i.e., A1 and c-IAP2 (42). These two biologic effects, i.e., anti-inflammatory and cytoprotective, are accomplished by different end products of heme degradation, i.e., labile Fe controls NF-B while CO is cytoprotective without interfering with NF-B (20).

In conclusion, the effects reported hereby provide clues as to our understanding of the protective effect of HO-1 in countering the pathogenesis of inflammatory diseases, including autoimmune neuroinflammation (63) as well as neuroinflammation emanating from unfettered responses to microbial challenge, e.g., cerebral malaria (62). Our present data support the notion that these protective effects are mediated, at least partially, via the ability of HO-1 to modulate EC labile Fe content, an effect that regulates NF-B activity in a manner that is anti-inflammatory and cytoprotective. Finally, these data might also help in providing a mechanism for the palliative effects of Fe chelation therapies.

	Acknowledgments

 
We thank John Eaton (University of Louisville, Louisville, KY) for instrumental help in setting-up the cellular iron assays, Rex Tyrrell (University of Bath, Bath, U.K.) for helpful discussion and for providing N-(2-hydroxybenzyl)-L-serine (HBSer), Prem Ponka (McGill University, Faculty of Medicine, Quebec, Canada) for helpful discussion and for providing salicylaldehyde isonicotinoylhydrazone (SIH), Federica Marelli-Berg and Anthony Dorling (Imperial College London, Hammersmith Hospital, London, U.K.) for help in setting-up primary mouse endothelial cell cultures, Nuno Moreno (Instituto Gulbenkian de Ciência, Oeiras, Portugal) for expert support with cell imaging, Sérgio Dias and Christophe Gregoire (Instituto Gulbenkian de Ciência) for reviewing the manuscript, Sofia Rebelo and Silvia Cardoso (Instituto Gulbenkian de Ciência) for invaluable support in maintenance of the Hmox1^–/– colony, and, finally, all other members of the "Inflammation Laboratory" (Instituto Gulbenkian de Ciência) for intellectual input.

	Disclosures

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
The authors have no financial conflict of interest.

	Footnotes

 
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

¹ This work was supported by the grants POCTI/BIA-BCM/56829/2004 and POCTI/SAU-MNO/56066/2004 (to M.S.), Ph.D. fellowships SFRH/BD/2990/2000 (to M.P.S.) and SFRH/BD/28016/2006 (to N.P.), postdoctoral fellowships SFRH/BPD/6303/2001 and SFRH/BPD/21072/2004 (to G.S.), SFRH/BPD/9380/2002 (to I.P.G.) and SFRH/BPD/25436/2005 (to R.L.), all from "Fundação para Ciência e a Tecnologia" Portugal, the European Commission’s Sixth Framework Programme contract no. LSHB-CT-2006-037377, Xenome (to M.P.S.). Research described in this article was supported in part by Philip Morris USA.

² M.P.S. and G.S. contributed equally to this work.

³ Address correspondence and reprint requests to Dr. Miguel Soares, Instituto Gulbenkian de Ciencia, Rua da Quinta Grande 6, Oeiras, 2780-156, Portugal. E-mail address: mpsoares{at}igc.gulbenkian.pt

⁴ Abbreviations used in this paper: EC, endothelial cell; BAEC, bovine aortic EC; CM-H₂DCFDA, 5-(and-6)-chloromethyl-2',7'-dichlorofluorescein diacetate, acetyl ester; DAPI, 4' 6-diamidino-2-phenylindole; DFO, deferoxamine mesylate; Fe, iron; HBSer, N-(2-hydroxybenzyl)-L-serine; HEK-293, human epithelial kidney 293 (cell); H-ferritin, heavy chain ferritin; HO-1, heme oxygenase 1; HPRT, hypoxanthine-guanine-phosphoribosyltransferase; luc, luciferase; MEC, mouse endothelial cell; MEF, mouse embryonic fibroblast; NAC, N-acetyl-L-cysteine; PKA, protein kinase A; RHD, Rel homology domain; ROS, reactive oxygen species; SIH, salicylaldehyde isonicotinoylhydrazone; TAD, transactivation domain; tet°, tetracycline operon; TET, tet° DNA binding domain; VP16, virion protein 16.

Received for publication January 24, 2007. Accepted for publication September 20, 2007.

	References

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 

1. Cines, D. B., E. S. Pollak, C. A. Buck, J. Loscalzo, G. A. Zimmerman, R. P. McEver, J. S. Pober, T. M. Wick, B. A. Konkle, B. S. Schwartz, et al 1998. Endothelial cells in physiology and in the pathophysiology of vascular disorders. Blood 91: 3527-3561. [Free Full Text]
2. Mantovani, A., F. Bussolino, M. Introna. 1997. Cytokine regulation of endothelial cell function: from molecular level to the bedside. Immunol. Today 18: 231-240. [Medline]
3. Wrighton, C. J., W. R. Hofer, T. Moll, R. Eytner, F. H. Bach, R. de Martin. 1996. Inhibition of endothelial cell activation by adenovirus-mediated expression of IB, an inhibitor of the transcription factor NF-B. J. Exp. Med. 183: 1013-1022. [Abstract/Free Full Text]
4. Collins, T., M. A. Read, A. S. Neish, M. Z. Whitley, D. Thanos, T. Maniatis. 1995. Transcriptional regulation of endothelial cell adhesion molecules: NF-B and cytokine-inducible enhancers. FASEB J. 9: 899-909. [Abstract]
5. Nabel, G. J., I. M. Verma. 1993. Proposed NF-B/I B family nomenclature. Genes Dev. 7: 2063[Free Full Text]
6. Schmitz, M. L., P. A. Baeuerle. 1991. The p65 subunit is responsible for the strong transcription activating potential of NF-B. EMBO J. 10: 3805-3817. [Medline]
7. Urban, M. B., R. Schreck, P. A. Baeuerle. 1991. NF-B contacts DNA by a heterodimer of the p50 and p65 subunit. EMBO J. 10: 1817-1825. [Medline]
8. Baeuerle, P. A., D. Baltimore. 1988. Activation of DNA-binding activity in an apparently cytoplasmic precursor of the NF-B transcription factor. Cell 53: 211-217. [Medline]
9. Baldwin, A. S., Jr. 1996. The NF-B and IB proteins: new discoveries and insights. Annu. Rev. Immunol. 14: 649-683. [Medline]
10. Chen, Z., J. Hagler, V. J. Palombella, F. Melandri, D. Scherer, D. Ballard, T. Maniatis. 1995. Signal-induced site-specific phosphorylation targets IB to the ubiquitin-proteasome pathway. Genes Dev. 9: 1586-1597. [Abstract/Free Full Text]
11. Chen, L. F., W. C. Greene. 2004. Shaping the nuclear action of NF-B. Nat. Rev. Mol. Cell Biol. 5: 392-401. [Medline]
12. Anrather, J., G. Racchumi, C. Iadecola. 2005. cis-Acting, element-specific transcriptional activity of differentially phosphorylated nuclear factor-B. J. Biol. Chem. 280: 244-252. [Abstract/Free Full Text]
13. Anrather, J., V. Csizmadia, M. P. Soares, H. Winkler. 1999. Regulation of NF-B RelA phosphorylation and transcriptional activity by p21^ras and protein kinase C in primary endothelial cells. J. Biol. Chem. 274: 13594-13603. [Abstract/Free Full Text]
14. Zhong, H., H. Suyang, H. Erdjument-Bromage, P. Tempst, S. Ghosh. 1997. The transcriptional activity of NF-B is regulated by the IB-associated PKAc subunit through a cyclic AMP-independent mechanism. Cell 89: 413-424. [Medline]
15. Vermeulen, L., G. De Wilde, P. Van Damme, W. Vanden Berghe, G. Haegeman. 2003. Transcriptional activation of the NF-B p65 subunit by mitogen- and stress-activated protein kinase-1 (MSK1). EMBO J. 22: 1313-1324. [Medline]
16. Duran, A., M. T. Diaz-Meco, J. Moscat. 2003. Essential role of RelA Ser311 phosphorylation by zetaPKC in NF-B transcriptional activation. EMBO J. 22: 3910-3918. [Medline]
17. Leung, T. H., A. Hoffmann, D. Baltimore. 2004. One nucleotide in a B site can determine cofactor specificity for NF-B dimers. Cell 118: 453-464. [Medline]
18. Bach, F. H., W. W. Hancock, C. Ferran. 1997. Protective genes expressed in endothelial cells: a regulatory response to injury. Immunol. Today 18: 483-486. [Medline]
19. Soares, M. P., Y. Lin, J. Anrather, E. Csizmadia, K. Takigami, K. Sato, S. T. Grey, R. B. Colvin, A. M. Choi, K. D. Poss, F. H. Bach. 1998. Expression of heme oxygenase-1 can determine cardiac xenograft survival. Nat. Med. 4: 1073-1077. [Medline]
20. Soares, M. P., M. P. Seldon, I. P. Gregoire, T. Vassilevskaia, P. O. Berberat, J. Yu, T. Y. Tsui, F. H. Bach. 2004. Heme oxygenase-1 modulates the expression of adhesion molecules associated with endothelial cell activation. J. Immunol. 172: 3553-3563. [Abstract/Free Full Text]
21. Ogawa, K., J. Sun, S. Taketani, O. Nakajima, C. Nishitani, S. Sassa, N. Hayashi, M. Yamamoto, S. Shibahara, H. Fujita, K. Igarashi. 2001. Heme mediates derepression of Maf recognition element through direct binding to transcription repressor Bach1. EMBO J. 20: 2835-2843. [Medline]
22. Alam, J., D. Stewart, C. Touchard, S. Boinapally, A. M. Choi, J. L. Cook. 1999. Nrf2, a Cap‘n’Collar transcription factor, regulates induction of the heme oxygenase-1 gene. J. Biol. Chem. 274: 26071-26078. [Abstract/Free Full Text]
23. Tenhunen, R., H. S. Marver, R. Schmid. 1968. The enzymatic conversion of heme to bilirubin by microsomal heme oxygenase. Proc. Natl. Acad. Sci. USA 61: 748-755. [Free Full Text]
24. Eisenstein, R. S., M. D. Garcia, W. Pettingell, H. N. Munro. 1991. Regulation of ferritin and heme oxygenase synthesis in rat fibroblasts by different forms of iron. Proc. Natl. Acad. Sci. USA 88: 688-692. [Abstract/Free Full Text]
25. Harrison, P. M., P. Arosio. 1996. Ferritins: molecular properties, iron storage function and cellular regulation. Biochim. Biophys. Acta 1275: 161-203. [Medline]
26. Cozzi, A., B. Corsi, S. Levi, P. Santambrogio, A. Albertini, P. Arosio. 2000. Overexpression of wild type and mutated human ferritin H-chain in HeLa cells: in vivo role of ferritin ferroxidase activity. J. Biol. Chem. 275: 25122-25129. [Abstract/Free Full Text]
27. Berberat, P. O., M. Katori, E. Kaczmarek, D. Anselmo, C. Lassman, B. Ke, X. Shen, R. W. Busuttil, K. Yamashita, E. Csizmadia, et al 2003. Heavy chain ferritin acts as an antiapoptotic gene that protects livers from ischemia reperfusion injury. FASEB J. 17: 1724-1726. [Abstract/Free Full Text]
28. Pham, C. G., C. Bubici, F. Zazzeroni, S. Papa, J. Jones, K. Alvarez, S. Jayawardena, E. De Smaele, R. Cong, C. Beaumont, et al 2004. Ferritin heavy chain upregulation by NF-B inhibits TNF-induced apoptosis by suppressing reactive oxygen species. Cell 119: 529-542. [Medline]
29. Ferris, C., S. Jaffrey, A. Sawa, M. Takahashi, S. Brady, R. Barrow, S. Tysoc, H. Wolosker, d. Baranano, S. Dore, et al 1999. Haem oxygenase-1 prevents cell death by regulating cellular iron. Nat. Cell Biol. 1: 152-157. [Medline]
30. Stocker, R., Y. Yamamoto, A. F. McDonagh, A. N. Glazer, B. N. Ames. 1987. Bilirubin is an antioxidant of possible physiological importance. Science 235: 1043-1046. [Abstract/Free Full Text]
31. Kutty, R. K., M. D. Maines. 1981. Purification and characterization of biliverdin reductase from rat liver. J. Biol. Chem. 256: 3956-3962. [Abstract/Free Full Text]
32. Yet, S. F., M. A. Perrella, M. D. Layne, C. M. Hsieh, K. Maemura, L. Kobzik, P. Wiesel, H. Christou, S. Kourembanas, M. E. Lee. 1999. Hypoxia induces severe right ventricular dilatation and infarction in heme oxygenase-1 null mice. J. Clin. Invest. 103: R23-R29. [Medline]
33. Marelli-Berg, F. M., E. Peek, E. A. Lidington, H. J. Stauss, R. I. Lechler. 2000. Isolation of endothelial cells from murine tissue. J. Immunol. Methods 244: 205-215. [Medline]
34. Beg, A. A., W. C. Sha, R. T. Bronson, S. Ghosh, D. Baltimore. 1995. Embryonic lethality and liver degeneration in mice lacking the RelA component of NF-B. Nature 376: 167-170. [Medline]
35. Laemmli, U. K.. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680-685. [Medline]
36. Qian, M. W., J. W. Eaton. 1991. Iron translocation by free fatty acids. Am. J. Pathol. 139: 1425-1434. [Abstract]
37. Artiss, J. D., S. Vinogradov, B. Zak. 1981. Spectrophotometric study of several sensitive reagents for serum iron. Clin. Biochem. 14: 311-315. [Medline]
38. Silva, G., A. Cunha, I. P. Gregoire, M. P. Seldon, M. P. Soares. 2006. The antiapoptotic effect of heme oxygenase-1 in endothelial cells involves the degradation of p38 MAPK isoform. J. Immunol. 177: 1894-1903. [Abstract/Free Full Text]
39. Livak, K. J., T. D. Schmittgen. 2001. Analysis of relative gene expression data using real-time quantitative PCR and the 2^–Ct method. Methods 25: 402-408. [Medline]
40. McDaid, J., K. Yamashita, A. Chora, R. Ollinger, T. B. Strom, X. C. Li, F. H. Bach, M. P. Soares. 2005. Heme oxygenase-1 modulates the allo-immune response by promoting activation-induced cell death of T cells. FASEB J. 19: 458-460. [Abstract/Free Full Text]
41. Brouard, S., L. E. Otterbein, J. Anrather, E. Tobiasch, F. H. Bach, A. M. Choi, M. P. Soares. 2000. Carbon monoxide generated by heme oxygenase 1 suppresses endothelial cell apoptosis. J. Exp. Med. 192: 1015-1026. [Abstract/Free Full Text]
42. Brouard, S., P. O. Berberat, E. Tobiasch, M. P. Seldon, F. H. Bach, M. P. Soares. 2002. Heme oxygenase-1-derived carbon monoxide requires the activation of transcription factor NF-B to protect endothelial cells from tumor necrosis factor--mediated apoptosis. J. Biol. Chem. 277: 17950-17961. [Abstract/Free Full Text]
43. Brostjan, C., J. Anrather, V. Csizmadia, G. Natarajan, H. Winkler. 1997. Glucocorticoids inhibit E-selectin expression by targeting NF-B and not ATF/c-Jun. J. Immunol. 158: 3836-3844. [Abstract]
44. Minami, T., W. C. Aird. 2001. Thrombin stimulation of the vascular cell adhesion molecule-1 promoter in endothelial cells is mediated by tandem nuclear factor-B and GATA motifs. J. Biol. Chem. 276: 47632-47641. [Abstract/Free Full Text]
45. Brostjan, C., J. Anrather, V. Csizmadia, G. Natarajan, F. H. Bach, H. Winkler. 1996. Deletion analysis of the porcine E-selectin promoter. Transplant. Proc. 28: 639-641. [Medline]
46. Madge, L. A., J. S. Pober. 2000. A phosphatidylinositol 3-kinase/Akt pathway, activated by tumor necrosis factor or interleukin-1, inhibits apoptosis but does not activate NFB in human endothelial cells. J. Biol. Chem. 275: 15458-15465. [Abstract/Free Full Text]
47. Momose, T., A. Tonegawa, J. Takeuchi, H. Ogawa, K. Umesono, K. Yasuda. 1999. Efficient targeting of gene expression in chick embryos by microelectroporation. Dev. Growth Differ. 41: 335-344. [Medline]
48. Kitazawa, M., K. Iwasaki. 1999. Reduction of ultraviolet light-induced oxidative stress by amino acid-based iron chelators. Biochim. Biophys. Acta 1473: 400-408. [Medline]
49. Creighton-Gutteridge, M., R. M. Tyrrell. 2002. A novel iron chelator that does not induce HIF-1 activity. Free Radical Biol. Med. 33: 356-363. [Medline]
50. Horackova, M., P. Ponka, Z. Byczko. 2000. The antioxidant effects of a novel iron chelator salicylaldehyde isonicotinoyl hydrazone in the prevention of H₂O₂ injury in adult cardiomyocytes. Cardiovasc. Res. 47: 529-536. [Abstract/Free Full Text]
51. Poss, K. D., S. Tonegawa. 1997. Heme oxygenase 1 is required for mammalian iron reutilization. Proc. Natl. Acad. Sci. USA 94: 10919-10924. [Abstract/Free Full Text]
52. Yachie, A., Y. Niida, T. Wada, N. Igarashi, H. Kaneda, T. Toma, K. Ohta, Y. Kasahara, S. Koizumi. 1999. Oxidative stress causes enhanced endothelial cell injury in human heme oxygenase-1 deficiency. J. Clin. Invest. 103: 129-135. [Medline]
53. Hentze, M. W., M. U. Muckenthaler, N. C. Andrews. 2004. Balancing acts: molecular control of mammalian iron metabolism. Cell 117: 285-297. [Medline]
54. Xiong, S., H. She, H. Takeuchi, B. Han, J. F. Engelhardt, C. H. Barton, E. Zandi, C. Giulivi, H. Tsukamoto. 2003. Signaling role of intracellular iron in NF-B activation. J. Biol. Chem. 278: 17646-17654. [Abstract/Free Full Text]
55. Anrather, J., V. Csizmadia, C. Brostjan, M. P. Soares, F. H. Bach, H. Winkler. 1997. Inhibition of bovine endothelial cell activation in vitro by regulated expression of a transdominant inhibitor of NF-B. J. Clin. Invest. 99: 763-772. [Medline]
56. Bowie, A. G., P. N. Moynagh, L. A. O’Neill. 1997. Lipid peroxidation is involved in the activation of NF-B by tumor necrosis factor but not interleukin-1 in the human endothelial cell line ECV304: lack of involvement of H₂O₂ in NF-B activation by either cytokine in both primary and transformed endothelial cells. J. Biol. Chem. 272: 25941-25950. [Abstract/Free Full Text]
57. Antosiewicz, J., W. Ziolkowski, J. J. Kaczor, A. Herman-Antosiewicz. 2007. Tumor necrosis factor--induced reactive oxygen species formation is mediated by JNK1-dependent ferritin degradation and elevation of labile iron pool. Free Radical Biol. Med. 43: 265-270. [Medline]
58. Zhong, H., M. J. May, E. Jimi, S. Ghosh. 2002. The phosphorylation status of nuclear NF-B determines its association with CBP/p300 or HDAC-1. Mol. Cell 9: 625-636. [Medline]
59. Ollivier, V., G. C. Parry, R. R. Cobb, D. de Prost, N. Mackman. 1996. Elevated cyclic AMP inhibits NF-B-mediated transcription in human monocytic cells and endothelial cells. J. Biol. Chem. 271: 20828-20835. [Abstract/Free Full Text]
60. Ryu, H., J. Lee, S. Impey, R. R. Ratan, R. J. Ferrante. 2005. Antioxidants modulate mitochondrial PKA and increase CREB binding to D-loop DNA of the mitochondrial genome in neurons. Proc. Natl. Acad. Sci. USA 102: 13915-13920. [Abstract/Free Full Text]
61. Otterbein, L. E., M. P. Soares, K. Yamashita, F. H. Bach. 2003. Heme oxygenase-1: unleashing the protective properties of heme. Trends Immunol. 24: 449-455. [Medline]
62. Pamplona, A., A. Ferreira, J. Balla, V. Jeney, G. Balla, S. Epiphanio, A. Chora, C. D. Rodrigues, I. P. Gregoire, M. Cunha-Rodrigues, et al 2007. Heme oxygenase-1 and carbon monoxide suppress the pathogenesis of experimental cerebral malaria. Nat. Med. 13: 703-710. [Medline]
63. Chora, A. A., P. Fontoura, A. Cunha, T. F. Pais, S. Cardoso, P. P. Ho, L. Y. Lee, R. A. Sobel, L. Steinman, M. P. Soares. 2007. Heme oxygenase-1 and carbon monoxide suppress autoimmune neuroinflammation. J. Clin. Invest. 117: 438-447. [Medline]

This article has been cited by other articles:

				G. Silva, V. Jeney, A. Chora, R. Larsen, J. Balla, and M. P. Soares Oxidized Hemoglobin Is an Endogenous Proinflammatory Agonist That Targets Vascular Endothelial Cells J. Biol. Chem., October 23, 2009; 284(43): 29582 - 29595. [Abstract] [Full Text] [PDF]

				R. Takamiya, C.-C. Hung, S. R. Hall, K. Fukunaga, T. Nagaishi, T. Maeno, C. Owen, A. A. Macias, L. E. Fredenburgh, A. Ishizaka, et al. High-Mobility Group Box 1 Contributes to Lethality of Endotoxemia in Heme Oxygenase-1-Deficient Mice Am. J. Respir. Cell Mol. Biol., August 1, 2009; 41(2): 129 - 135. [Abstract] [Full Text] [PDF]

				A. R. Kinderlerer, I. Pombo Gregoire, S. S. Hamdulay, F. Ali, R. Steinberg, G. Silva, N. Ali, B. Wang, D. O. Haskard, M. P. Soares, et al. Heme oxygenase-1 expression enhances vascular endothelial resistance to complement-mediated injury through induction of decay-accelerating factor: a role for increased bilirubin and ferritin Blood, February 12, 2009; 113(7): 1598 - 1607. [Abstract] [Full Text] [PDF]

				S. Naidu, N. Wijayanti, S. Santoso, T. Kietzmann, and S. Immenschuh An Atypical NF-{kappa}B-Regulated Pathway Mediates Phorbol Ester-Dependent Heme Oxygenase-1 Gene Activation in Monocytes J. Immunol., September 15, 2008; 181(6): 4113 - 4123. [Abstract] [Full Text] [PDF]

				C. Hu, A. Dandapat, L. Sun, J. Chen, M. R. Marwali, F. Romeo, T. Sawamura, and J. L. Mehta LOX-1 deletion decreases collagen accumulation in atherosclerotic plaque in low-density lipoprotein receptor knockout mice fed a high-cholesterol diet Cardiovasc Res, July 15, 2008; 79(2): 287 - 293. [Abstract] [Full Text] [PDF]

				X. Li, M. G. Schwacha, I. H. Chaudry, and M. A. Choudhry Heme Oxygenase-1 Protects against Neutrophil-Mediated Intestinal Damage by Down-Regulation of Neutrophil p47phox and p67phox Activity and O2- Production in a Two-Hit Model of Alcohol Intoxication and Burn Injury J. Immunol., May 15, 2008; 180(10): 6933 - 6940. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Seldon, M. P. Articles by Soares, M. P. Search for Related Content PubMed PubMed Citation Articles by Seldon, M. P. Articles by Soares, M. P.