Intracellular metal ion chelators inhibit TNFα-induced SP-1 activation and adhesion molecule expression in human aortic endothelial cells

Free Radical Biology & Medicine, Vol. 34, No. 6, pp. 674 – 682, 2003 Copyright © 2003 Elsevier Science Inc. Printed in the USA. All rights reserved 0891-5849/03/$–see front matter

doi:10.1016/S0891-5849(02)01375-8

Original Contribution INTRACELLULAR METAL ION CHELATORS INHIBIT TNF␣-INDUCED SP-1 ACTIVATION AND ADHESION MOLECULE EXPRESSION IN HUMAN AORTIC ENDOTHELIAL CELLS WEI-JIAN ZHANG

and

BALZ FREI

Linus Pauling Institute, Oregon State University, Corvallis, OR, USA (Received 9 July 2002; Revised 20 November 2002; Accepted 3 December 2002)

Abstract—Endothelial adhesion molecule expression and monocyte recruitment are causal events in human atherosclerosis, and are believed to be caused, in part, by oxidative stress. Because redox-active transition metal ions, such as iron and copper, play an essential role in the generation of free radicals and the initiation and propagation of lipid peroxidation, we hypothesized that transition metal ions may also be involved in endothelial activation. Therefore, we investigated the effects of the intracellular iron-chelator, desferrioxamine (DFO), and the intracellular copper-chelator, neocuproine (NC), on TNF␣-induced expression of adhesion molecules in human aortic endothelial cells (HAEC). Treatment of HAEC with DFO (0.01– 0.1 mM) or NC (0.1 and 0.5 mM) time- and dose-dependently inhibited TNF␣-induced protein expression of E-selectin, vascular cell adhesion molecule-1 (VCAM-1), and intercellular adhesion molecule-1 (ICAM-1). In contrast, iron-saturated DFO and the extracellular copper chelator, bathocuproinedisulfonic acid, had no effect on adhesion molecule expression. DFO and NC also dose-dependently inhibited TNF␣induced upregulation of adhesion molecule mRNA levels. Furthermore, treatment of HAEC with 0.5 mM DFO or NC completely inhibited TNF␣-induced activation of the transcription factor, specificity protein-1 (SP-1), but only partially inhibited or did not affect activation of other transcription factors known to regulate adhesion molecule expression, i.e., nuclear factor ␬B (NF␬B), activator protein-1 (AP-1), and interferon regulatory factor-1 (IRF-1). Finally, inhibition of endothelial nitric oxide synthase with N-nitro-L-arginine methylester (0.5 mM) did not attenuate the inhibitory effects of the metal ion chelators on adhesion molecule expression. Our data suggest that intracellular, but not extracellular, transition metal ions mediate inflammatory cytokine-induced SP-1 activation and adhesion molecule expression in endothelial cells. © 2003 Elsevier Science Inc. Keywords—Metal ions, Cell adhesion molecules, Endothelium, Signal transduction, Atherosclerosis

INTRODUCTION

endotoxin, and some reactive oxygen species (ROS) [2]. Studies in humans have found increased adhesion molecule expression on prelesion sites and atherosclerotic lesions [3,4]. In mice, genetic deficiencies of VCAM-1 or ICAM-1 are associated with decreased atherosclerosis [5–7]. Therefore, modulation of adhesion molecule expression and monocyte-endothelial interactions may be important targets for the prevention and treatment of atherosclerotic vascular disease. There is some, albeit inconsistent, evidence that iron plays a role in the pathogenesis of atherosclerosis. The association between body iron status and the risk of cardiovascular diseases (CVD) was first postulated by Sullivan [8], and is supported by some epidemiological studies [9,10]. Catalytically active iron has been detected in early and advanced atherosclerotic lesions [11]. Some

Adhesion of circulating monocytes to activated endothelium and their subsequent transendothelial migration are important steps in the initiation and progression of atherosclerosis [1]. These processes depend on the coordinated expression of adhesion molecules on endothelial cells, including vascular cell adhesion molecule-1 (VCAM-1), intercellular adhesion molecule-1 (ICAM1), and E-selectin [2]. The inappropriate expression of endothelial adhesion molecules is induced in response to “injury” by various inflammatory cytokines, bacterial Address correspondence to: Balz Frei, Ph.D., Linus Pauling Institute, Oregon State University, 571 Weniger Hall, Corvallis, OR 97331, USA; Tel: (541) 737-5075; Fax: (541) 737-5077; E-Mail: [email protected]. 674

Metal ion chelation inhibits endothelial activation

animal experiments have found that the severity of atherosclerosis is either enhanced by iron overload [12] or reduced by iron deficiency [13], while others have shown that iron overload decreases atherosclerosis in conjunction with hypo- or hypercholesterolemic effects [14,15]. In vitro studies have demonstrated that transition metal ions are required for LDL oxidation by cultured endothelial cells [16], macrophages [17], and smooth muscle cells [18]. Increased copper body status also has been suggested to be an independent risk factor for ischemic heart disease and CVD mortality [19,20]. Copper has been detected in advanced human atherosclerotic lesions [21], and a recent study in mice has shown that copperinduced arterial injury generates lesions reminiscent of those seen in human atherosclerosis [22]. Because redoxactive copper is capable of promoting lipid peroxidation and generating ROS in cells, it is conceivable that copper plays a role in atherogenesis through its effects on endothelial cells, leukocytes, and platelets, in addition to affecting lipoprotein metabolism and oxidation [23]. To test the hypothesis that transition metal ions play a role in endothelial activation, we used desferrioxamine (DFO), a ferric iron chelator that is clinically used in the treatment of transfusional iron overload [24], and neocuproine (NC), a lipophilic copper (I)-specific chelator that has been used to inhibit copper-mediated oxidative damage in vitro and in vivo [25,26]. The effects of these metal ion chelators on TNF␣-induced expression of cellular adhesion molecules were investigated in human aortic endothelial cells (HAEC). Because the transcription factors, nuclear factor ␬ B (NF␬B), activator protein-1 (AP-1), specificity protein-1 (SP-1), and interferon regulatory factor-1 (IRF-1), all have binding sites in the promoter region of the genes of VCAM-1 and other adhesion molecules [27–29], we also examined the effect of metal ion chelators on TNF␣-induced activation of these transcription factors. MATERIALS AND METHODS

Materials

675

media (Clonetics) at 37°C in a humidified 95% air-5% CO2 atmosphere. For experiments, cells were grown to confluence in 75 cm2 flasks, petri dishes (100 mm) or 96 well plates (Costar, Cambridge, MA, USA) coated with 1% calf skin gelatin (Sigma), using endothelial culture medium (ECM) consisting of M199 medium (Sigma) supplemented with 20% fetal calf serum (GIBCO, Grand Island, NY, USA), 100 ng/ml streptomycin, 100 IU/ml penicillin, 250 ng/ml fungizone, 1 mM glutamine (GIBCO), and 1 ng/ml human recombinant basic fibroblast growth factor (Boehringer Mannheim). In all experiments HAEC of passage seven or eight were used. Experiments HAEC were incubated for 16 h with DFO or NC, or for 1 h with PDTC. Control cells were incubated for the same periods of time with either media alone (for DFO and PDTC) or media containing the same concentration (ⱕ 0.1%) of the vehicle DMSO (for NC). Subsequently, the cells were washed twice with M199 medium and coincubated with the same concentrations of DFO, NC, or PDTC and TNF␣ (10 U/ml) for various time periods, viz.: 4.5 h for protein expression of adhesion molecules; 2 h for measurement of mRNA levels; and 1 h for measurement of activation of transcription factors. In some experiments, the cells were incubated for 1 or 16 h, respectively, with the nitric oxide (NO) synthase inhibitor, L-NAME (0.5 mM), or the protein synthesis inhibitor cycloheximide (1 ␮ g/ml) with or without DFO or NC, before addition of TNF␣ and incubation for another 2 or 4.5 h. Cell viability Cell viability was assessed by morphology using phase contrast microscopy and by reduction of the tetrazolium salt MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) by mitochondrial dehydrogenases, according to the manufacturer’s instructions (Boehringer Mannheim).

Desferrioxamine, neocuproine (2,9-dimethyl-1,10phenanthroline), pyrrolidine dithiocarbamate (PDTC), bathocuproinedisulfonic acid, N-nitro-L-arginine methylester (L-NAME), and cycloheximide were obtained from Sigma (St. Louis, MO, USA). Human recombinant TNF␣ was purchased from Boehringer Mannheim Biochemica (Indianapolis, IN, USA). Other biochemicals used are mentioned below and were of the highest purity commercially available.

Measurement of cell adhesion molecules by ELISA

Culture of endothelial cells

Northern blot analysis

HAEC were obtained from Clonetics (San Diego, CA, USA) and cultured with endothelial cell growth

Total cellular RNA was isolated from HAEC using TRIzol Reagent (Life Technologies Inc., Grand Island,

Surface expression of adhesion molecules (E-selectin, VCAM-1, and ICAM-1) was quantified by a cell ELISA as described [30]. The antibodies used for the assay were mouse monoclonal anti-human E-selectin (R & D Systems, Minneapolis, MN, USA), ICAM-1 (R & D Systems), or VCAM-1 (DAKO, Carpinteria, CA, USA).

676

W.-J. ZHANG and B. FREI

Fig. 1. Desferrioxamine, but not iron-loaded desferrioxamine, dose-dependently inhibits TNF␣-induced protein expression of Eselectin, VCAM-1, and ICAM-1. Human aortic endothelial cells were incubated for 16 h without or with different concentrations (0.01, 0.05, or 0.1 mM) of DFO or DFO loaded with equimolar amounts of ferric iron (Fe-DFO). The cells were then washed twice with M199 medium and incubated for 4.5 h with TNF␣ (10 U/ml) and the same concentrations of DFO or Fe-DFO. Surface expression of adhesion molecules was measured by ELISA, and results were calculated as percentage of TNF␣ stimulation as described in Materials and Methods. Data shown are the mean values ⫾ SD of four separate incubations and are representative of three independent experiments. *p ⬍ .05 compared with TNF␣ alone.

NY, USA). Northern blot analysis was performed as described [31]. RNA blots were hybridized with 106 cpm/ml of the [␣-32P]dATP-labeled oligonucleotide probes for either human E-selectin, VCAM-1, or ICAM-1 (R & D Systems) overnight at 57°C in a hybridization oven. Blots were washed, air dried, and exposed to Hyperfilm X-ray films (Amersham, Piscataway, NJ, USA) at ⫺80°C. The equal loading of total RNA was confirmed by staining of 28S and 18S ribosomal RNA with ethidium bromide. Electrophoretic mobility shift assay (EMSA) Nuclear extracts were prepared and EMSA was performed as described [30]. Briefly, 5 ␮g of nuclear extract was incubated for 30 min at room temperature with either 32P-labeled double-stranded NF␬B oligonucleotide, 5'-AGTTGAGGGGACTTTCCCAGGC-3'; AP-1 oligonucleotide, 5'-CGCTTGATGACTCAGCCGGAA-3'; SP-1 oligonucleotide, 5'-ATTCGATCGGGGCGGGGCGAGC-3; or IRF-1 oligonucleotide, 5'-GGAAGCGAAAATGAAATTGACT (Santa Cruz Biotechnology, Santa Cruz, CA, USA). The DNA-protein complex was electrophoresed on a 5% nondenaturing polyacrylamide gel in 0.5 ⫻ Tris/borate EDTA buffer. The specificity of binding was examined by incubating nuclear extracts prepared from TNF␣-treated cells for 30 min at room temperature with 2– 4 ␮g of antibodies against the p65 or p50 subunit of

NF␬B; the c-Jun or c-Fos subunit of AP-1; SP-1; or IRF-1 (Santa Cruz Biotechnology) before addition of the 32Plabeled oligonucleotide. Specificity was further confirmed by competition with a 100-fold excess of unlabeled competitor consensus oligonucleotide. Radioactive bands were detected by autoradiography at ⫺80°C. Statistical analysis Data are reported as mean ⫾ SD. One-way ANOVA or student’s unpaired t-test were used for statistical analysis of the original data, and significance was accepted at the p ⬍ .05 level. Data in Figs. 1 and 2 are expressed as “% of TNF␣ stimulation,” which was calculated as follows: (value for TNF␣ plus DFO- or NC-treated cells/ value for TNF␣-treated cells) ⫻ 100%. Percent inhibition was calculated as follows: [1 ⫺ (value for TNF␣ plus DFO- or NC-treated cells ⫺ value for unstimulated cells/value for TNF␣-treated cells ⫺ value for unstimulated cells)] ⫻ 100%. RESULTS

Intracellular metal ion chelators inhibit TNF␣-induced protein expression of endothelial adhesion molecules To investigate the role of intracellular metal ions in adhesion molecule expression, HAEC were treated for

Metal ion chelation inhibits endothelial activation

677

Fig. 2. Neocuproine dose-dependently inhibits TNF␣-induced protein expression of E-selectin, VCAM-1, and ICAM-1. Human aortic endothelial cells were incubated for 16 h without (negative control) or with NC (0.1 or 0.5 mM), or for 1 h with PDTC (0.05 or 0.1 mM; positive control). The cells were then washed twice with M199 medium and incubated for 4.5 h with TNF␣ (10 U/ml) and the same concentrations of NC or PDTC. Surface expression of adhesion molecules was measured by ELISA, and results were calculated as percentage of TNF␣ stimulation as described in Materials and Methods. Data shown are the mean values ⫾ SD of four separate incubations and are representative of three independent experiments. *p ⬍ .05 compared with TNF␣ alone.

16 h with DFO (0.01, 0.05, and 0.1 mM) or NC (0.1 and 0.5 mM), followed by incubation with TNF␣ (10 U/ml) for 4.5 h. As shown in Figs. 1 and 2, respectively, treatment of cells with DFO or NC dose-dependently inhibited TNF␣-induced expression of E-selectin, VCAM-1, and ICAM-1. For example, E-selectin, VCAM-1, and ICAM-1 protein levels were reduced by 74 ⫾ 2.5%, 84 ⫾ 3.3%, and 59 ⫾ 1.7%, respectively, by 0.1 mM DFO (P ⬍ .01 compared with TNF␣ control; n ⫽ 3) (Fig. 1); and 93 ⫾ 3.6%, 84 ⫾ 3.0%, and 89 ⫾ 4.0%, respectively, by 0.5 mM NC (P ⬍ .01; n ⫽ 3) (Fig. 2). PDTC (0.05 and 0.1 mM), used in our experiments as a “positive” control, also dose-dependently inhibited TNF␣-induced adhesion molecule expression in HAEC (Fig. 2), confirming previous data [32]. Importantly, DFO was ineffective at inhibiting adhesion molecule expression when it was preloaded with Fe3⫹ (Fig. 1). These data strongly suggest that the inhibitory effect of DFO on adhesion molecule expression is due to metal ion chelation, and not the known ROS scavenging activity of DFO [33] or adaptation of the cells in response to the presence of DFO. In addition, when cells were co-incubated with DFO (0.1 mM) and TNF␣, or treated for only 1 h with NC (0.1 mM) prior to TNF␣ addition, or treated with the membrane-impermeable copper chelator, bathocuproinedisulfonic acid (0.5 mM), adhesion molecule expression was not inhibited

(data not shown). In all the above experiments, the cells remained fully viable and no cytotoxic effects were observed, as assessed by microscopic examination of cell morphology and the MTT assay (data not shown). As it has been shown that nitric oxide (NO) can inhibit adhesion molecule expression mediated by NF␬B [34], we further investigated whether the above inhibitory effects of metal ion chelators are mediated by increased NO levels. To this end, HAEC were incubated for 1 h with L-NAME (0.5 mM), a competitive inhibitor of endothelial NO synthase. However, this treatment affected neither TNF␣-induced adhesion molecule expression nor the inhibitory effects of DFO or NC (data not shown). Intracellular metal ion chelators inhibit TNF␣-induced upregulation of adhesion molecule mRNA levels To further define the mechanism by which DFO and NC inhibit adhesion molecule expression, mRNA levels of E-selectin, VCAM-1, and ICAM-1 were determined. As shown in Fig. 3, incubation of HAEC for 2 h with 10 U/ml TNF␣ resulted in strong induction of message levels for all three adhesion molecules. Treatment of the cells with DFO (0.1 and 0.5 mM) or NC (0.1 and 0.5 mM) for 16 h significantly inhibited TNF␣-induced mRNA expression in a dose-dependent manner (Fig. 3A

678

W.-J. ZHANG and B. FREI

Fig. 3. Desferrioxamine and neocuproine dose-dependently inhibit TNF␣-induced upregulation of E-selectin, VCAM-1, and ICAM-1 mRNA levels. Human aortic endothelial cells were incubated for 16 h without (negative control) or with different concentrations (0.1 or 0.5 mM) of (A) DFO or (B) NC, or (C) for 1 h with PDTC (0.01 or 0.1 mM; positive control). The cells were then washed twice with M199 medium and incubated for 2 h with TNF␣ (10 U/ml) and the same concentrations of DFO, NC, or PDTC. Northern blot analysis for adhesion molecules was performed as described in Materials and Methods. The equal loading of total RNA was confirmed by staining of 28S and 18S ribosomal RNA with ethidium bromide. Data shown are representative of two experiments.

and 3B). As determined by densitometric analysis of the Northern blots, E-selectin, VCAM-1, and ICAM-1 message levels were decreased by 90, 100, and 64%, respectively, by 0.1 mM DFO (n ⫽ 2) (Fig. 3A); and 90, 94, and 49% by 0.5 mM NC (n ⫽ 2) (Fig. 3B). As expected [30], treatment of cells with PDTC (0.01 and 0.1 mM) also dose-dependently inhibited TNF␣-induced upregulation of mRNA levels of the three adhesion molecules (Fig. 3C). Interestingly, addition of cycloheximide (1 ␮g/ml) abolished the inhibitory effects of NC, but not DFO, on adhesion molecule mRNA expression (Fig. 4). In fact, NC seemed to obliterate the inhibitory effect of cycloheximide itself on TNF␣-induced mRNA expression (Fig. 4). These data indicate that protein synthesis is required for NC to inhibit adhesion molecule expression, in contrast to DFO. Intracellular metal ion chelators only partially inhibit TNF␣-induced NF␬B activation, but completely inhibit SP-1 activation As activation of the transcription factor NF␬B is required for the transcriptional induction of endothelial adhesion molecules [27–29], we evaluated the effects of DFO and NC on TNF␣-induced NF␬B activation, using EMSA. As shown in Fig. 5, treatment of HAEC with

Fig. 4. Cycloheximide abolishes the inhibitory effects of neocuproine, but not desferrioxamine, on TNF␣-induced upregulation of adhesion molecule mRNA levels. Human aortic endothelial cells were incubated for 16 h without or with DFO (0.5 mM) or NC (0.5 mM) and cycloheximide (1 ␮g/ml). The cells were then washed twice with M199 medium and incubated for 2 h with TNF␣ (10 U/ml) and the same concentrations of DFO or NC with or without cycloheximide. Northern blot analysis for adhesion molecules was performed as described in Materials and Methods. The equal loading of total RNA was confirmed by staining of 28S and 18S ribosomal RNA with ethidium bromide. Data shown are representative of two experiments.

Metal ion chelation inhibits endothelial activation

679

Fig. 5. Desferrioxamine and neocuproine only partially inhibit TNF␣-induced NF␬B activation. Human aortic endothelial cells were incubated for 16 h without (negative control) or with DFO (0.5 mM) or NC (0.5 mM), or for 1 h with PDTC (0.1 mM; positive control). The cells were then washed twice with M199 medium and incubated for 1 h with TNF␣ (10 U/ml) and the same concentrations of DFO, NC, or PDTC. Nuclear extracts were isolated and EMSA was performed as described in Materials and Methods. Control lanes included antibodies against the p65 (␣-p65) or p50 (␣-p50) subunit of NF␬B, or a 100-fold excess of unlabeled NF␬B consensus oligonucleotide (cold probe). Data shown are representative of two experiments.

DFO (0.5 mM) or NC (0.5 mM) for 16 h inhibited TNF␣-stimulated NF␬B DNA binding activity by 31 and 55%, respectively, as quantitated by densitometry. However, these same concentrations of the metal ion chelators almost completely inhibited adhesion molecule expression (see, e.g., Fig. 3). In contrast, PDTC (0.1 mM) abolished TNF␣-induced NF␬B binding activity (Fig. 5), in agreement with its effect on adhesion molecule expression. These data suggest that other transcription factors may play a role in the inhibitory effects of metal ion chelators on adhesion molecule expression. Specifically, AP-1, SP-1, and IRF-1 are known to have consensus binding sequences in the 5'-flanking region of the human VCAM-1 gene [28,29]. Interestingly, we found that treatment of HAEC with DFO (0.5 mM) or NC (0.5 mM) completely inhibited TNF␣-stimulated SP-1 DNA binding activity (Fig. 6). In contrast, TNF␣-induced AP-1 and IRF-1 activation was not inhibited by DFO, and partially inhibited by about 60% and 55%, respectively, by NC (Fig. 6). In addition to abolishing NF␬B activation (see Fig. 5), PDTC (0.1 mM) also completely inhibited SP-1 and IRF-1 activation, but did not inhibit AP-1 activation (data not shown).

DISCUSSION

This study demonstrates for the first time that the intracellular metal ion chelators, desferrioxamine and neocuproine, effectively inhibit TNF␣-induced activa-

tion of endothelial cells, as manifested by decreased adhesion molecule expression. In these experiments, DFO acted as a metal ion chelator, not by scavenging

Fig. 6. Desferrioxamine and neocuproine completely inhibit TNF␣induced SP-1 activation, but not AP-1 or IRF-1 activation. Human aortic endothelial cells were incubated for 16 h without (negative control) or with DFO (0.5 mM) or NC (0.5 mM). The cells were then washed twice with M199 medium and incubated for 1 h with TNF␣ (10 U/ml) and the same concentrations of DFO or NC. Nuclear extracts were isolated and EMSA was performed as described in Materials and Methods. Control lanes included a 100-fold excess of unlabeled SP-1, AP-1, or IRF-1 consensus oligonucleotide (cold probe). Data shown are representative of two experiments.

680

W.-J. ZHANG and B. FREI

ROS [33], because prior loading of DFO with equimolar amounts of ferric iron abolished its inhibitory effect on adhesion molecule expression. Furthermore, the effect was limited to chelation of intracellular metal ions, as supported by two lines of evidence: (i) chelation of metal ions in the cell culture medium with the extracellular copper chelator, bathocuproinedisulfonic acid, did not inhibit adhesion molecule expression; and (ii) when the cells were co-incubated with DFO and TNF␣ or treated for only 1 h with NC prior to TNF␣ addition—not enough time for DFO or NC to effectively penetrate the cell membrane and chelate intracellular metal ions—no inhibition of adhesion molecule expression was observed. It is well known that ROS production and LDL oxidation by vascular cells are mediated by transition metal ions [16,18,35], and that the oxidative capacity of these cells is greatly enhanced by micromolar concentrations of iron or copper [18,36]. Therefore, our data suggest that chelation of intracellular metal ions decreases ROS generation and oxidative stress, and thus inhibits endothelial activation. In agreement with this notion, DFO has been shown to inhibit iron-mediated production of ROS [33] and protect against ROS-induced myocardial and coronary endothelial ischemia-reperfusion injury [37,38]. Furthermore, DFO has been shown to ameliorate endothelial dysfunction by improving endothelium-dependent vasodilation in patients with coronary artery disease [39]. The inhibitory effect of NC on endothelial adhesion molecule expression strongly implicates intracellular copper, as NC is not only membrane-permeable but also chelates copper(I) with high affinity and specificity [25]. Intracellular copper exists in several oxidation states, and can change from one redox state to another under physiological conditions. This redox cycling is considered to be responsible for both the physiological functions of copper and its toxicity [35]. Interestingly, it has been shown recently that there is less than one free intracellular copper ion per cell [40], suggesting that the observed inhibitory effect of NC may be due to removal of copper from (a) protein(s). In contrast, free redox-active iron can be detected in cells [35], and its chelation by DFO or 1,10-phenanthroline can inhibit H2O2-induced DNA damage [41]. Therefore, the observed effects of DFO on adhesion molecule expression are likely mediated by chelation of iron from this intracellular labile pool. Different mechanisms of action for NC and DFO are further suggested by our findings that protein synthesis is required for the inhibition of adhesion molecule expression by NC, but not DFO. It is tempting to speculate that this difference in dependence on protein synthesis is related to the chelation of protein-bound copper vs. free iron by NC and DFO, respectively.

One possible mechanism that we considered for the inhibitory effects of NC and DFO on endothelial activation is inhibition of the NF␬B signaling pathway. NF␬B, a transcription factor that regulates expression of many inflammatory genes, has been proposed to be redoxsensitive [42]. In most cell types, NF␬B can be activated by a diverse range of stimuli, suggesting that several signaling pathways are involved. Lipid peroxidation, but not H2O2, has been reported to play a role in TNF␣induced NF␬B activation in EC304 cells [43]. Redoxactive transition metal ions likely play a key role in the initiation and propagation of lipid peroxidation, leading to the generation of peroxyl and alkoxyl radicals, as well as lipid hydroperoxides [35]. Consistently, DFO and PDTC, which both strongly inhibit iron-dependent lipid peroxidation, also strongly inhibited TNF␣-induced NF␬B activation in ECV304 cells [43]. However, in the present study, we found that DFO and NC only partially inhibited NF␬B activation in HAEC. In agreement with this observation, it has been shown that various antioxidants such as ␣-tocopherol, nordihydroguatiaretic acid, and the NADH/NADPH oxidase inhibitor, diphenylene iodonium, inhibit cytokineinduced adhesion molecule expression in endothelial cells without also inhibiting NF␬B activation [44 – 46]. Therefore, we considered other transcription factors that have binding sites in the promoter region of the VCAM-1 and other adhesion molecule genes [28,29]. Among these transcription factors, SP-1, but not AP-1 or IRF-1, was inhibited by DFO and NC to the same degree that these metal ion chelators inhibited transcriptional activation of adhesion molecules. Interestingly, SP-1 has been suggested to be activated by oxidative stress, and was the only transcription factor involved in upregulation of VCAM-1 gene expression by interleukin-4 in human umbilical vein endothelial cells [47]. In addition or alternatively to promoting lipid peroxidation, transition metal ions may stimulate adhesion molecule expression by inhibiting the biological activity of NO. Nitric oxide, an iron ligand produced by endothelial NO synthase, has been shown to inhibit adhesion molecule expression in endothelial cells, most likely via inhibition of NF␬B [34]. In contrast, our data indicate that NO is not involved in TNF␣-induced HAEC activation and its inhibition by DFO or NC, as the NO synthase inhibitor, L-NAME, was without effect. This result is also consistent with the finding that SP-1 is more important than NF␬B in the inhibitory effects of metal ion chelators on adhesion molecule expression, as SP-1 activity is not known to be inhibited by NO. In conclusion, the protective effects of desferrioxamine and neocuproine demonstrated in this study shed new light on our understanding of the role of intracellular, redox-active transition metal ions and the transcription

Metal ion chelation inhibits endothelial activation

factor, specificity protein-1, in vascular endothelial activation by inflammatory cytokines. Further studies are necessary to fully understand the underlying mechanisms and to investigate whether metal ion chelators may also be useful in the prevention and treatment of atherosclerosis and other inflammatory conditions in vivo.

[17]

[18]

[19] Acknowledgements — This work was supported by grants ES-11542 (W.J.Z.) and HL-60886 (B.F.) from the U.S. National Institutes of Health, and a Pilot Project Grant (W.J.Z.) from the Linus Pauling Institute, Oregon State University.

REFERENCES [1] Ross, R. Atherosclerosis—an inflammatory disease. N. Engl. J. Med. 340:115–126; 1999. [2] Carlos, T. M.; Harlan, J. M. Leukocyte-endothelial adhesion molecules. Blood 84:2068 –2101; 1994. [3] Cybulsky, M. I.; Gimbrone, M. A. Jr. Endothelial expression of a mononuclear leukocyte adhesion molecule during atherogenesis. Science 251:788 –791; 1991. [4] Van der Wal, A. C.; Das, P. K.; Tigges, A. J.; Becker, A. E. Adhesion molecules on the endothelium and mononuclear cells in human atherosclerotic lesions. Am. J. Pathol. 141:1427–1433; 1992. [5] Cybulsky, M. I.; Iiyama, K.; Li, H.; Zhu, S.; Chen, M.; Iiyama, M.; Davis, V.; Gutierrez-Ramos, J. C.; Connelly, P. W.; Milstone, D. S. A major role for VCAM-1, but not ICAM-1, in early atherosclerosis. J. Clin. Invest. 107:1255–1262; 2001. [6] Collins, R. G.; Velji, R.; Guevara, N. V.; Hicks, M. J.; Chan, L.; Beaudet, A. L. P-selectin or intercellular adhesion molecule (ICAM)-1 deficiency substantially protects against atherosclerosis in apolipoprotein E-deficient mice. J. Exp. Med. 191:189 –194; 2000. [7] Nageh, M. F.; Sandberg, E. T.; Marotti, K. R.; Lin, A. H.; Melchior, E. P.; Bullard, D. C.; Beaudet, A. L. Deficiency of inflammatory cell adhesion molecules protects against atherosclerosis in mice. Arterioscler. Thromb. Vasc. Biol. 17:1517–1520; 1997. [8] Sullivan, J. L. Iron and sex difference in heart disease risk. Lancet 1:1293–1294; 1981. [9] Salonen, J. T.; Nyyssonen, K.; Korpela, H.; Tuomilehto, J.; Seppanen, R.; Salonen, R. High stored iron levels are associated with excess risk of myocardial infarction in eastern Finnish men. Circulation 86:803– 811; 1992. [10] Kiechl, S.; Willeit, J.; Egger, G.; Poewe, W.; Oberhollenzer, F. Body iron stores and the risk of carotid atherosclerosis: prospective results from the Bruneck study. Circulation 96:3300 –3307; 1997. [11] Pang, J. H.; Jiang, M. J.; Chen, Y. L.; Wang, F. W.; Wang, D. L.; Chu, S. H.; Chau, L. Y. Increased ferritin gene expression in atherosclerotic lesions. J. Clin. Invest. 97:2204 –2212; 1996. [12] Araujo, J. A.; Romano, E. L.; Brito, B. E.; Parthe, V.; Romano, M.; Bracho, M.; Montano, R. F.; Cardier, J. Iron overload augments the development of atherosclerotic lesions in rabbits. Arterioscler. Thromb. Vasc. Biol. 15:1172–1180; 1995. [13] Lee, T. S.; Shiao, M. S.; Pan, C. C.; Chau, L. Y. Iron-deficient diet reduces atherosclerotic lesions in apoE-deficient mice. Circulation 99:1222–1229; 1999. [14] Dabbagh, A. J.; Shwaery, G. T.; Keaney, J. F. Jr.; Frei, B. Effect of iron overload and iron deficiency on atherosclerosis in the hypercholesterolemic rabbit. Arterioscler. Thromb. Vasc. Biol. 17:2638 –2645; 1997. [15] Kirk, E. A.; Heinecke, J. W.; LeBoeuf, R. C. Iron overload diminishes atherosclerosis in apoE-deficient mice. J. Clin. Invest. 107:1545–1553; 2001. [16] Martin, A.; Frei, B. Both intracellular and extracellular vitamin C inhibit atherogenic modification of LDL by human vascular en-

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

681

dothelial cells. Arterioscler. Thromb. Vasc. Biol. 17:1583–1590; 1997. Leake, D. S.; Rankin, S. M. The oxidative modification of lowdensity lipoproteins by macrophages. Biochem. J. 270:741–748; 1990. Heinecke, J. W.; Rosen, H.; Chait, A. Iron and copper promote modification of low density lipoprotein by human arterial smooth muscle cells in culture. J. Clin. Invest. 74:1890 –1894; 1984. Kok, F. J.; Van Duijn, C. M.; Hofman, A.; Van der Voet, G. B.; De Wolff, F. A.; Paays, C. H.; Valkenburg, H. A. Serum copper and zinc and the risk of death from cancer and cardiovascular disease. Am. J. Epidemiol. 128:352–359; 1988. Salonen, J. T.; Salonen, R.; Korpela, H.; Suntioinen, S.; Tuomilehto, J. Serum copper and the risk of acute myocardial infarction: a prospective population study in men in eastern Finland. Am. J. Epidemiol. 134:268 –276; 1991. Swain, J.; Gutteridge, J. M. Prooxidant iron and copper, with ferroxidase and xanthine oxidase activities in human atherosclerotic material. FEBS Lett. 368:513–515; 1995. Ploplis, V. A.; Cornelissen, I.; Sandoval-Cooper, M. J.; Weeks, L.; Noria, F. A.; Castellino, F. J. Remodeling of the vessel wall after copper-induced injury is highly attenuated in mice with a total deficiency of plasminogen activator inhibitor-1. Am. J. Pathol. 158:107–117; 2001. Ferns, G. A.; Lamb, D. J.; Taylor, A. The possible role of copper ions in atherogenesis: the Blue Janus. Atherosclerosis 133:139 – 152; 1997. Klaassen, C. D. Heavy metals and heavy-metal antagonist. In: Gilman, A. G.; Goodman, L. S.; Gilman, A., eds. The pharmacological basis of therapeutics. New York: Macmillan; 1985: 1605–1627. de Mello Filho, A. C.; Meneghini, R. Protection of mammalian cells by o-phenanthroline from lethal and DNA-damaging effects produced by active oxygen species. Biochim. Biophys. Acta 847: 82– 89; 1985. Applebaum, Y. J.; Kuvin, J.; Borman, J. B.; Uretzky, G.; Chevion, M. The protective role of neocuproine against cardiac damage in isolated perfused rat hearts. Free Radic. Biol. Med. 8:133–143; 1990. Collins, T.; Read, M. A.; Neish, A. S.; Whitley, M. Z.; Thanos, D.; Maniatis, T. Transcriptional regulation of endothelial cell adhesion molecules: NF-kappa B and cytokine-inducible enhancers. FASEB J. 9:899 –909; 1995. Neish, A. S.; Williams, A. J.; Palmer, H. J.; Whitley, M. Z.; Collins, T. Functional analysis of the human vascular cell adhesion molecule 1 promoter. J. Exp. Med. 176:1583–1593; 1992. Neish, A. S.; Khachigian, L. M.; Park, A.; Baichwal, V. R.; Collins, T. Sp1 is a component of the cytokine-inducible enhancer in the promoter of vascular cell adhesion molecule-1. J. Biol. Chem. 270:28903–28909; 1995. Zhang, W. J.; Frei, B. Alpha-lipoic acid inhibits TNF-alphainduced NF-kappaB activation and adhesion molecule expression in human aortic endothelial cells. FASEB J. 15:2423–2432; 2001. Zhang, W. J.; Wojta, J.; Binder, B. R. Notoginsenoside R1 counteracts endotoxin-induced activation of endothelial cells in vitro and endotoxin-induced lethality in mice in vivo. Arterioscler. Thromb. Vasc. Biol. 17:465– 474; 1997. Marui, N.; Offermann, M. K.; Swerlick, R.; Kunsch, C.; Rosen, C. A.; Ahmad, M.; Alexander, R. W.; Medford, R. M. Vascular cell adhesion molecule-1 (VCAM-1) gene transcription and expression are regulated through an antioxidant-sensitive mechanism in human vascular endothelial cells. J. Clin. Invest. 92: 1866 –1874; 1993. Halliwell, B. Use of desferrioxamine as a ‘probe’ for iron-dependent formation of hydroxyl radicals. Evidence for a direct reaction between desferal and the superoxide radical. Biochem. Pharmacol. 34:229 –233; 1985. De Caterina, R.; Libby, P.; Peng, H. B.; Thannickal, V. J.; Rajavashisth, T. B.; Gimbrone, M. A. Jr.; Shin, W. S.; Liao, J. K. Nitric oxide decreases cytokine-induced endothelial activation. J. Clin. Invest. 96:60 – 68; 1995.

682

W.-J. ZHANG and B. FREI

[35] Halliwell, B.; Gutteridge, J. M. C. Role of free radicals and catalytic metal ions in human disease: an overview. Methods Enzymol. 186:1– 85; 1990. [36] Morel, D. W.; DiCorleto, P. E.; Chisolm, G. M. Endothelial and smooth muscle cells alter low density lipoprotein in vitro by free radical oxidation. Arteriosclerosis 4:357–364; 1984. [37] Bolli, R.; Jeroudi, M. O.; Patel, P. S.; Aruoma, O. I.; Halliwell, B.; Lai, E. K.; McCay, P. B. Marked reduction of free radical generation and contractile dysfunction by antioxidant therapy begun at the time of reperfusion: evidence that myocardial ‘stunning’ is a manifestation of reperfusion injury. Circ. Res. 65:607– 622; 1989. [38] Sellke, F. W.; Shafique, T.; Ely, D. L.; Weintraub, R. M. Coronary endothelial injury after cardiopulmonary bypass and ischemic cardioplegia is mediated by oxygen-derived free radicals. Circulation 88:II395–II400; 1993. [39] Duffy, S. J.; Biegelsen, E. S.; Holbrook, M.; Russell, J. D.; Gokce, N.; Keaney, J. F. Jr.; Vita, J. A. Iron chelation improves endothelial function in patients with coronary artery disease. Circulation 103:2799 –2804; 2001. [40] Rae, T. D.; Schmidt, P. J.; Pufahl, R. A.; Culotta, V. C.; O’Halloran, T. V. Undetectable intracellular free copper: the requirement of a copper chaperone for superoxide dismutase. Science 284:805– 808; 1999. [41] Barbouti, A; Doulias, P. T.; Zhu, B. Z.; Frei, B.; Galaris, D. Intracellular iron, but not copper, plays a critical role in hydrogen

[42] [43]

[44] [45]

[46]

[47]

peroxide-induced DNA damage. Free Radic. Biol. Med. 31:490 – 498; 2001. Schreck, R.; Albermann, K.; Baeuerle, P. A. Nuclear factor B: an oxidative stress-responsive transcription factor of eukaryotic cells. Free Radic. Res. Commun. 17:221–237; 1992. Bowie, A. G.; Moynagh, P. N.; O’Neill, L. A. Lipid peroxidation is involved in the activation of NF-kappaB by tumor necrosis factor but not interleukin-1 in the human endothelial cell line ECV304. Lack of involvement of H2O2 in NF-kappaB activation by either cytokine in both primary and transformed endothelial cells. J. Biol. Chem. 272:25941–25950; 1997. Faruqi, R.; de la Motte, C.; DiCorleto, P. E. Alpha-tocopherol inhibits agonist-induced monocytic cell adhesion to cultured human endothelial cells. J. Clin. Invest. 94:592– 600; 1994. Ahmad, M.; Theofanidis, P.; Medford, R. M. Role of activating protein-1 in the regulation of the vascular cell adhesion molecule-1 gene expression by tumor necrosis factor-alpha. J. Biol. Chem. 273:4616 – 4621; 1998. Tummala, P. E.; Chen, X. L.; Medford, R. M. NF-kappa B independent suppression of endothelial vascular cell adhesion molecule-1 and intercellular adhesion molecule-1 gene expression by inhibition of flavin binding proteins and superoxide production. J. Mol. Cell. Cardiol. 32:1499 –1508; 2000. Lee, Y. W.; Kuhn, H.; Hennig, B.; Neish, A. S.; Toborek, M. IL-4-induced oxidative stress upregulates VCAM-1 gene expression in human endothelial cells. J. Mol. Cell. Cardiol. 33:83–94; 2001.