reoxygenating microvascular endothelium exhibits temporal dissociation of nf-κb and ap-1 activation

Original Contribution

REOXYGENATING MICROVASCULAR ENDOTHELIUM EXHIBITSTEMPORAL DISSOCIATION OF NF-�B AND AP-1 ACTIVATION

RAMESH NATARAJAN, BERNARD J. FISHER, DREW G. JONES, SHOBHA GHOSH, and ALPHA A. FOWLER III*Center for Vascular Inflammation Research, Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine,

Virginia Commonwealth University, Richmond, VA, USA

(Received 4 October 2001; Accepted 7 March 2002)

Abstract—Alterations of cellular redox balance in microvascular endothelium results in changes of essential cellfunctions. These alterations may arise, in part, due to modifications in the pattern of gene expression produced bytranscription factor activation. Endothelium subjected to hypoxia/reoxygenation becomes redox imbalanced, therebyleading to activation and perhaps production of a proinflammatory state. A human dermal microvascular endothelial cellline (HMEC-1) was exposed to 6 h of hypoxia (3% O2) followed by return to normoxia atmospheric conditions. Reactiveoxygen species (ROS) generation (dichlorofluoroscein epifluorescence) was immediate and significant followingreoxygenation. Electrophoretic mobility shift assays revealed activation of the oxidant sensitive transcription factorsNF�B and AP-1, though importantly, peak activation of each factor was separated temporally by greater than 60 min.NF�B activation occurred without degradation of the inhibitory protein I�B�. Reoxygenating HMEC-1 exhibited agreater than 500-fold increase in polymorphonuclear neutrophil (PMN) adhesion when compared to normoxic controls.Exposure of reoxygenating HMEC-1 to the antioxidant pyrrolidine dithiocarbamate produced complete abrogation ofNF�B activation and the intensive PMN adhesion observed in untreated, posthypoxic HMEC-1. Though rexoygenationstress induced significant upregulation of PMN adhesion, no upregulation of interleukin-8 production was observed. Ourresults suggest that ROS generation occurring in endothelium following onset of reoxygenation stress signals activationof key transcription factors and that their activation takes place in a temporal fashion. The temporal feature oftranscription factor activation may be key to production of a postischemic proinflammatory state. © 2002 ElsevierScience Inc.

Keywords—Reactive oxygen species, Endothelium, NF�B, AP-1, Hypoxia, Reoxygenation, Neutrophil adhesionInterleukin-8, Free radicals

INTRODUCTION

Periods of ischemic injury to vascular beds followed byrestoration of perfusion commonly results in significantorgan damage with ensuing morbid or even fatal com-plications [1–4]. A significant body of converging re-search suggests that parenchymal microvascular endo-thelium and activated polymorphonuclear neutrophils(PMN) play essential roles in precipitating tissue damageassociated with reperfusion injury [5]. Reactive oxygen

species (ROS) generated within microvascular endothe-lium following onset of reoxygenation promotes tran-scription and expression of key adhesion molecules (e.g.,E-selectin, Intercellular adhesion molecule-1 [ICAM-1])and release of potent proinflammatory chemokines, (e.g.,interleukin-8 [IL-8]) which in turn coordinate PMN ad-hesion and transvascular migration [6,7]. Recent re-search suggests that microvascular endothelium plays acentral role in modulating PMN adhesion and activation.Menger and colleagues and Steinbauer et al. found rapidonset of PMN rolling followed by striking increases intethered PMN adhesion during early reperfusion injuryof skeletal muscle [8,9]. Topham and colleagues reportedthat production of the CXC chemokine interleukin-8 bycytokine-activated endothelium was crucial to the signal-ing events that lead to PMN activation and tissue injury

Drew G. Jones is recipient of the Virginia Thoracic Society ResearchFellowship Training Award.

Address correspondence to: Alpha A. Fowler, III, M.D., Professor ofMedicine, Chairman, Division of Pulmonary and Critical Care Medi-cine, Department of Internal Medicine, Box 980050, Virginia Com-monwealth University, Richmond, VA 23298, USA; Tel: (804) 828-3558; Fax: (804) 828-3559; E-Mail: [email protected].

Free Radical Biology & Medicine, Vol. 32, No. 10, pp. 1033–1045, 2002Copyright © 2002 Elsevier Science Inc.Printed in the USA. All rights reserved

0891-5849/02/$–see front matter

PII S0891-5849(02)00813-4

1033

[10]. Endothelium activated by reoxygenation to produceIL-8 may have a similar impact on PMN activity [11,12].The biological events that ensue following endothelialcell activation and PMN adhesion results in altered mi-crovascular permeability and prompt migration of acti-vated PMN, ultimately producing injury and organ dys-function [13].

Increasingly ROS, at concentrations substantially be-low that required for oxidative damage, are recognized ascritical for cell signaling and regulation of gene expres-sion [14,15]. The “oxidant sensitive” transcription fac-tors nuclear factor-�B (NF-�B) and activator protein-1(AP-1) have assumed critical importance as knowledgein this area has expanded. NF-�B alone or in concertwith AP-1 regulates genes central to inflammation (e.g.,interleukin-6, ICAM-1, interleukin-8, tumor necrosisfactor-�) [16]. ROS generated in the cytosol and mito-chondria are central to regulation of the redox state ofprotein cysteinyl residues, and therefore, form a commonmechanism for regulating protein conformation andfunction [17]. In the current study, a well-characterizedhuman microvascular endothelial cell line was employedto examine key physiological and molecular events pro-duced by reoxygenation stress, occurring following ex-posure to acute hypoxia. Our intent with these studieswas to use the modeling system described here to gainnew insights into the mechanisms that lead to endothelialcell-derived vascular inflammation.

MATERIALS AND METHODS

Reagents and kits

The Interleukin-8 ELISA kit was obtained from Bio-Source International (Camarillo, CA, USA). Polyacryl-amide Readygels were obtained from Biorad (Hercules,CA, USA). Electrophoretic mobility shift kits were ob-tained from Promega (Madison, WI, USA). Sterile tissueculture plasticware was obtained from Corning (Corning,NY, USA). Culture media was obtained from GIBCO-Invitrogen (Carlsbad, CA, USA). Specialty gases wereobtained from BOC gases (Murray Hill, NJ, USA). Thekit for determining cell viability (Live/Dead) was ob-tained from Molecular Probes (Eugene, OR, USA). Flu-orescent dyes: 2�,7�-dichlorodihydrofluorescein diacetate(H2DCF-DA) and Calcein AM were obtained from Mo-lecular Probes. The polyclonal antibody to Inhibitorkappa B alpha (sc-371) and oligonucleotide probes forNF-�B and AP-1 were obtained from Santa Cruz Bio-technology (Santa Cruz, CA, USA). Hypoxia chambers(Modular Incubator Chamber) were obtained from Bil-lups-Rothenberg Incorporated (Del Mar, CA, USA). Im-mobilon P membranes were obtained from Millipore(Bedford, MA, USA). All other chemicals and reagents

were obtained from Sigma Chemicals (St. Louis, MO,USA).

Endothelial cell culture

Endothelial cells utilized for this study were obtainedfrom the Center for Disease Control and Prevention(CDC), Atlanta, GA, USA. The cell line resulted fromtransfection of human dermal microvascular endothelialcells with a PBR-322-based plasmid containing the cod-ing region for the SV40 A gene product and large Tantigen. The cell line was immortalized by Dr. EdwinAdes, Mr. Fransisco J. Candal of the CDC, and Dr.Thomas Lawly of Emory University, and is designatedHMEC-1 [18]. Endothelial cells were cultured understerile conditions and maintained in medium MCDB-131supplemented with 10% fetal bovine serum, hydrocorti-sone (1 �g/ml), and epithelial cell growth factor (10ng/ml) under a 5% CO2 atmosphere, at 37°C.

Establishing hypoxia/reoxygenation conditions

Hypoxia was initiated by sealing confluent HMEC-1cultures within Billups-Rothenberg chambers. Chamberswere suffused for 10 min with a gas mixture consistingof 1% O2, 5% CO2, 94% N2 at a flow of 15 l/min.Chamber integrity was assessed in preliminary studies byanalysis of culture media for oxygen partial pressuresunder the conditions described. Repeated measures re-vealed that within 30 min of initiating hypoxia, mediaoxygen concentrations diminished to 3% and remainedunchanged for periods of up to 24 h. In this report,HMEC-1 cultures were exposed to hypoxic environ-ments for a period of 6 h. Preliminary studies showed thepartial pressures of oxygen in culture medium underhypoxic conditions to fall to a nadir of 30 � 3 mm Hgwithin 30 min. Fresh culture medium added at the outsetof reoxygenation possessed partial pressures of oxygenof 150 � 4 mm Hg. At 6 h, reoxygenation stress wasinitiated by opening chambers and returning cells toatmospheric oxygen tensions without additional oxygensupplemenation. At the initiation of reoxygenation, freshculture medium was exchanged and reoxygenating cellsincubated at 37°C in 5% CO2, and 95% air.

Fluorescence microscopy

Fluorescence imaging of HMEC-1 in this study wasperformed using an Olympus model IX70 inverted phasemicroscope (Olympus America, Melville, NY, USA)outfitted with an IX-FLA fluorescence observation sys-tem at a 10� magnification through an Uplan FI objec-tive (Olympus). Fluorescence images were digitized and

1034 R. NATARAJAN et al.

captured in real time by a MagnaFire digital camera(Optronics, Goleta, CA, USA). Output from the digitalcamera was directed by interface electronics to a DellDimension 4100 desktop computer running Image-Proimage analysis software (Media Cybernetics, SilverSpring, MD, USA) running under a Windows 98 envi-ronment (Microsoft, Redmond, WA, USA). Quantifica-tion of fluorescent cellular images from individual ex-periments was accomplished using a counting operationscommand contained within the Image-Pro software.

Assessing HMEC-1 viability during reoxygenationstress

To examine the extent to which reoxygenation stressaltered HMEC-1 viability, an assay that assesses viabilityor cytotoxicity based upon a two-color fluorescence(LIVE/DEAD, Molecular Probes) was used. The assayutilizes two probes (calcein AM and ethidium ho-modimer-1) that measure intracellular esterase activityand plasma membrane integrity; two recognized param-eters of cell viability [19]. Assays were performed bysimultaneously adding 150 �l of calcein AM (0.5 �M)and ethidium homodimer (1 �M) into wells containingreoxygenating cells. During the reoxygenation phase,cell-wells were incubated for 30 min at 22°C with theprobes. Calcein AM is retained within live cells, produc-ing an intense uniform green fluorescence whileethidium homodimer-1 enters dead cells having damagedmembranes. Damaged cells undergo a 40-fold enhance-ment of red fluorescence as ethidium homodimer-1 bindsto nucleic acids. Cells were imaged as described employ-ing both FITC and TRITC filters, respectively, for cal-cein AM and ethidium homodimer-1. ConfluentHMEC-1 cultures were subjected to 6 h of hypoxia.Medium was exchanged and reoxygenating cultures ex-amined at 1, 4, 6, and 12 h. Cells fluorescing green or redwere captured by fluorescence microscopy as describedand quantified via Image-Pro software. Our studies showthat HMEC-1 subjected to reoxygenation stress as de-scribed maintained greater than 99% viability.

Assessing oxidative stress in reoxygenating HMEC-1

The oxidant-sensing probe H2DCF-DA was used todetect intracellular oxidant stress during reoxygenation[20]. A 10 mM stock solution of H2DCF-DA was pre-pared in ethanol on the day of experimentation. Conflu-ent HMEC-1 undergoing reoxygenation stress and nor-moxic HMEC-1 controls were loaded with 10 �MH2DCF-DA via a 30 min incubation at 37°C. Cells werewashed thrice with warm PBS and returned to MCDBImedium. H2DCF-DA loading of HMEC-1 was accom-

plished prior to establishing experimental conditions.Normoxic and reoxygenating HMEC-1 were fixed andROS detected using an Olympus microscope as de-scribed above following excitation at 485 nm and emis-sion at 530 nm (Molecular Probes). This strategy per-mitted immediate imaging of ROS generation. Inaddition, HMEC-1 grown to confluence in 48 well plateswere loaded with H2DCF-DA and treated as describedabove. At 0 and 120 min posthypoxia, cells were washedand resuspended in PBS (250 �l per well). Relativefluorescence was measured at 485 ex/530 em using amicroplate fluorometer (Molecular Devices SPECTRA-max GeminiXS). Results are expressed in relative fluo-rescence units (RFU).

Quantification of interleukin-8 protein

Output of IL-8 protein was quantified in mediumobtained from cultures of reoxygenating HMEC-1 usinga sandwich ELISA prepared with human IL-8 Cytosetsantibody pair (Biosource International) according tomanufacturer’s instructions. Briefly, Maxisorb microtiterELISA plates (Nunc) were coated with mouse monoclo-nal anti-human IL-8 antibody (1 �g/ml, 100 �l per well)and stored overnight at 4°C. Plates were washed � 4 andblocked for 2 h at 22°C with 0.5% bovine serum albumin(300 �l per well). Plates were again washed � 4 beforeuse. A standard curve was prepared using serial dilutionsof recombinant human IL-8 (12.5–800 pg/ml). Standardsor samples were incubated (100 �l) with biotinylatedmouse monoclonal anti-human IL-8 antibody (50 �l) for2 h at 22°C with continual shaking (700 rpm). Plateswere washed � 4 and incubated with streptavidin-horse-radish peroxidase (100 �l) for 30 min at room tempera-ture. Plates were again washed � 4 and incubated withstabilized chromagen (hydrogen peroxide and tetrameth-ylbenzidine, 100 �l) for 15–30 min at 22°C in the dark.The reaction was stopped using 1.8 N sulfuric acid (50�l). Absorbance at 450 nm was determined and IL-8concentrations determined from a four parameter logisticcurve fit algorithm (Molecular Devices Softmax Pro).Results are expressed as pg/ml of IL-8 in HMEC-1supernates.

Inhibitor kappa B alpha Western blot analysis

Reoxygenating or control, normoxic HMEC-1 in 35mm dishes were rinsed once with ice-cold PBS and lysisbuffer (1X PBS, 1% Nonidet P-40, 0.5% deoxycholate,0.1% SDS) added. Cell lysates were passed through a#21 gauge needle, and centrifuged for 20 min at14,000 � g. Total cell lysates (10 �g protein) and lowmolecular weight markers were resolved by SDS poly-

1035Reoxygenation stress in HMEC-1

acrylamide gel electrophoresis (10%) and electrophoreti-cally transferred to polyvinylidene fluoride membranes(0.45 �m pore size). Immunodetection was performedusing a primary I�B� antibody (Santa Cruz Biotechnol-ogy) and the Renaissance Western Blot Chemilumines-cence Reagent Plus (NEN Life Science Products, Bos-ton, MA, USA).

Electrophoretic mobility shift assay (EMSA)

Reoxygenating or normoxic control HMEC-1 werewashed once (PBS), and exposed to buffer A (10 mMHEPES, 10 mM KCl, 100 �M EDTA, 100 �M EGTA,2 mM NaF, 2 mM Na3VO4, 100 �M PMSF) for 15 minat 4°C. Cells were policed into 1.5 ml tubes, lysed(Igepal CA-630, 0.05%), and centrifuged (13,000 � g).Nuclear pellets were resuspended in buffer B (20 mMHEPES, 390 mM NaCl, 1 mM EDTA, 1 mM EGTA, 2mM NaF, 2 mM Na3VO4, 2 mM PMSF), protein con-centration assessed and samples stored at �70°C [21,22]. EMSA was performed as described by Harant et al.[23]. Five �g of nuclear protein was incubated in abinding reaction with [32P] end-labeled double-strandedNF-�B oligonucleotide. DNA-binding reactions wereperformed at room temperature for 20 min in 50 mM Tris(pH 7.5), 50 mM NaCl, 1 mM dithiothreitol, 0.05%Igepal, and 2 �g poly d[I-C]. Samples were resolved on6% polyacrylamide gels at 100 V and imaged by auto-radiography. Specific NF-�B binding was verified by acompetitive EMSA with 100-fold excess unlabeled dou-ble-stranded NF-�B oligonucleotide in the binding reac-tion prior to the addition of the [32P] end-labeled oligo-nucleotide.

Neutrophil adhesion assay

Reoxygenation-induced changes in PMN adhesion toHMEC-1 were assessed using a modification of themethod of Braut-Boucher et al. [24]. Briefly, confluentHMEC-1 cells in 48 well plates were subjected to hyp-oxia/reoxygenation as described. Following informedconsent, using a protocol approved through the VirginiaCommonwealth University Health System InstitutionalReview Board, PMN were isolated from human blood asdescribed by Fowler et al. using density gradient centrif-ugation and hypotonic lysis [25] and adjusted to 5 � 106

cells/ml in MCDB-131. PMN were then incubated withCalcein-AM (5 �M) for 30 min at 22°C and washedthrice in PBS. Fifty thousand Calcein-AM labeled PMNwere then added to each well of plates that containedeither reoxygenating or normoxic HMEC-1. At the com-pletion of a prescribed incubation time all wells werewashed thrice with warm PBS. Calcein-AM labeled ad-

herent PMN remaining in each well were imaged andenumerated at 10� magnification using fluorescence mi-croscopy as described. For each condition, the numbersof adherent cells present in triplicate wells is reported.The results are expressed as cells per high-powered field.

Immunofluorescent imaging for E-selectin inposthypoxic HMEC-1

HMEC-1 were grown to confluence on fibronectin-coated glass cover slips. Cells were exposed to hypoxia(3% O2) for 6 h then re-oxygenation stress initiated byopening chambers and returning cells to atmosphericoxygen as described above. Cells were fixed in 3.7%formaldehyde/PBS for 15 min at 22°C, washed withPBS � 3, and blocked [PBS with 5% nonfat dry milk,1% normal chicken serum (NCS) and 0.1% triton X-100]for 40 min. Cells were incubated overnight at 4°C withE-selectin antibody (5 �g/ml) (Murine anti-human E-selectin, R&D Systems, BBA16) or murine anti-humanICAM-1 antibody (Santa Cruz Biotechnology, sc-107).Cells were washed � 3 with PBS and incubated withAlexa Fluor 488 anti-mouse antibody (10 �g/ml) (Mo-lecular Probes, A-21200) for 60 min at 22°C. Cells wereimaged as described above using an Olympus IX70 witha FITC filter and 60� oil immersion objective. Fluores-cence images were digitalized and captured by Mag-nafire digital camera. Studies were controlled with anonspecific murine IgG antibody.

Statistical analysis

Mean values were calculated from data obtained fromthree or more separate experiments and reported asmean � SEM. The significance of the difference be-tween groups with multiple comparisons was assessed byone-way analysis of variance (ANOVA) and by Student-Newman-Keuls test. Statistical significance was con-firmed at a p value � .05. A minimum of three indepen-dent experiments was used to confirm observations.

RESULTS

Reoxygenating HMEC-1 exhibit immediate andsignificant generation of oxidative stress

The onset of oxidant stress following initiation ofreoxygenation was immediate and significant. Figure 1Bshows a 10� magnification obtained within 60 s of theonset of reoxygenation, revealing significant epifluores-cence. In contradistinction, control cells imaged at anidentical magnification (Fig. 1A), exhibited little or noepifluorescence. Twenty-� (20�) magnification ofreoxygenating HMEC-1 revealed a homogeneous epiflu-


orescence evident within the cytoplasm immediately(�1 min) following onset of reoxygenation with thegreatest degree of epifluorescence observed in mitochon-dria positioned in a perinuclear location (Fig. 1D). Iden-tical studies utilizing MitoTracker Red fluorochome con-firmed this DCF perinuclear epifluorescence to residewithin mitochondria (data not shown). HMEC-1 cells

undergoing hypoxia/reoxygenation were also preincu-bated (prior to hypoxia) with the antioxidant pyrrolidinedithiocarbamate (PDTC, [10–50 �M]). As seen in Fig.1C, PDTC (50 �M) suppressed the production of ROS atthe onset of reoxygenation. Spectrofluorometeric quan-tification of oxidant stress generated by HMEC-1 duringreoxygenation stress (Fig. 1E) revealed that ROS gener-

Fig. 1. Detection of reactive oxygen species in reoxygenating endothelium. HMEC-1 cells were preloaded with 10 �M H2DCF-DAas described in Materials and Methods. They were exposed to hypoxic conditions (3% O2) for 6 h in the presence (C) or absence ofthe antioxidant PDTC followed by reoxygenation at room air (21% O2) for 1 min (A, B, D). Control cells were maintained under roomair conditions (A). Normoxic and reoxygenating HMEC-1 were fixed and ROS detected using an Olympus inverted fluorescencemicroscope following excitation at 485 nm and emission at 530 nm as described in Materials and Methods. (E) Spectrofluorometricquantification of ROS production. Posthypoxic HMEC-1 showed significantly greater (p � .05 vs. normoxic control) ROS productionat 0 min and significantly less (p � .05 vs. normoxic control) at 1 h. Posthypoxic HMEC-1 exhibited a significant rebound in ROSproduction at 2 h (#p � .05 vs. 60 min posthypoxic values). Posthypoxic HMEC-1 ROS production was not significantly different fromnormoxic controls for the remaining 10 h.


ation occurred immediately upon reoxygenation (� 1min) and was significantly greater than normoxic cells(p � .05). Oxidant stress in posthypoxic HMEC-1 fell tovalues significantly lower than normoxic controls by 1 h(p � .05) but rebounded significantly by 2 h to normoxiccontrol values. For the remaining 10 h of observation,oxidant stress in posthypoxic HMEC-1 was not signifi-cantly different from normoxic controls.

Reoxygenation stress in HMEC-1 promotesnonsimultaneous activation of NF-�B and AP-1

Electrophoretic mobility shift assays assessing activa-tion of NF-�B and AP-1 were performed on cultures ofreoxygenating HMEC-1. Following hypoxia/reoxygen-ation, cells were harvested at 0, 1, 2, 4, 6, 8, and 12 h andnuclear and cytoplasmic extracts purified as describedabove. Figure 2A shows the results of a representativeexperiment where serial assessments of NF-�B and AP-1activation were performed. Our studies revealed a greaterthan 5-fold increase in NF-�B activation in HMEC-1 at60 min. When compared to normoxic controls, thesechanges in NF-�B activation 60 min after the onset ofreoxygenation were highly significant (p � .001). By120 min following onset of reoxygenation stress, NF�Bactivation had attenuated to levels observed in normoxiccontrols. For the balance of the experimental period(2–12 h), the levels of nuclear NF�B protein did notchange in the nuclei of reoxygenating cells, remainingsimilar to controls. AP-1 activation as assessed byEMSA was less intense and followed a different kineticprofile than that observed for NF�B. Peak AP-1 activa-tion occurred at 120 min following onset of reoxygen-ation (1.9-fold increase), a time when NF�B protein inreoxygenating nuclei had returned to normoxic controllevels. Thus, our studies indicate that abrupt (within 60min) activation of NF-�B occurs following onset ofreoxygenation but lacking additional stresses return tothose levels observed in control cells. In further studies,NF�B activation was assessed in reoxygenatingHMEC-1 exposed to PDTC 10, 50, and 100 �M. PDTCsignificantly suppressed NF�B activation at every con-centration, underscoring the oxidant sensitive activationof NF�B (Fig. 2B).

NF-�B activation observed in reoxygenating HMEC-1occurs without I�B� degradation

Figure 3 shows a representative Western blot forI�B� examining cytoplasmic extracts obtained fromidentical cells reported in previously discussed EMSAexperiments. Our studies show that exposing HMEC-1 toreoxygenation stress does not induce significant degra-

dation of I�B� protein as compared to I�B� degradationobserved following cytokine exposure (tumor necrosisfactor-�, [TNF-�]). After incubating HMEC-1 withTNF-�, immediate I�B� breakdown (15 min) was ob-served followed by transcription of new I�B� protein.Thus, these data suggest that NF-�B activation occurringin the setting of significant oxidative activity exhibitsmarked stimulus specificity with IkB� degradation oc-curring following cytokine exposure but not followingoxidant stress produced in posthypoxic cells.

Reoxygenation stress promotes significant increases inneutrophil adhesion

Forty-eight well culture plates containing confluentHMEC-1 were subjected to 6 h of hypoxia. Mediumoverlying both hypoxic and normoxic HMEC-1 cultureswas removed and replaced with fresh complete MCDB-131 medium. All cultures were then incubated at 37°C,in 21% oxygen, 5% CO2, balance N2 for 120 min. At 120min 5 � 105 Calcein AM-labeled PMN were added to allwells. Following 10, 20, 30, and 45 min contact time, allwells were washed � 3 and the numbers of adherentPMN quantified. Figure 4 shows that for all contact timesexamined, reoxygenating HMEC-1 promoted signifi-cantly greater (p � .05) PMN adhesion than normoxiccontrols. In further experiments, the antioxidant PDTC(10–50 �M) was incubated with reoxygenating or nor-moxic HMEC-1. PMN adhesion was then determined inan identical fashion. Our results show that PDTC at allconcentrations studied abrogated PMN adhesion to post-hypoxic HMEC-1 (Fig. 5).

HMEC-1 subjected to reoxygenation stress do notproduce significantly greater IL-8 than normoxiccontrols

Conditioned medium obtained from reoxygenatingHMEC-1 was harvested and analyzed for IL-8 protein.Figure 6A shows IL-8 protein output by reoxygenatingHMEC-1 compared to that observed in normoxic con-trols. In these studies, the extent of IL-8 secretion byreoxygenating cells was not different than that observedby normoxic control cells over an 8 h observation period.Both cell populations exhibited a gradual accumulationof IL-8 in culture medium suggesting ongoing constitu-tive production. In additional studies HMEC-1 were in-cubated with TNF� (0.1–100 ng/ml). Figure 6B showsthat cytokine stress induced by TNF� promoted signifi-cant (p � .001) IL-8 production at every concentrationtested when compared to media controls.


Posthypoxic HMEC-1 exhibit significant upregulationof E-selectin and ICAM-1

Identical hypoxia/reoxygenation protocols to thatoutlined above for PMN adhesion studies were em-ployed to study the adhesion molecules E-selectin andICAM-1 during the initial phases of HMEC-1 rexoy-

genation. These studies were controlled with normoxicHMEC-1. The results are presented in Fig. 7. Ourstudies show a dramatic upregulation of E-selectin andICAM-1 at 2 h following onset of reoxygenation stress(Fig. 7B, D). In contradistinction, normoxic HMEC-1revealed little or no immunofluorescence, indicatingthat normoxic cells exhibit little constitutive E-selec-

Fig. 2. Activation of NF�B and AP-1 during reoxygenation stress. (A) HMEC-1 cells were maintained under normoxic conditions or exposedto hypoxic conditions (3% O2) for 6 h followed by reoxygenation at room air (21% O2) for the indicated times. Cells were harvested forpreparation of nuclear extracts. Electrophoretic mobility shift assay was performed using labeled double stranded consensus NF�B or AP-1oligonucleotides as described in Materials and Methods. A normoxic control (*) is shown in lane 1 of each representative assay. The relativeintensities of the shifted band were quantitated by densitometric analysis and expressed as percent normoxic control. Values are mean � SEMof three independent observations (#p � .001). (B) HMEC-1 cells were maintained under normoxic conditions or exposed to hypoxicconditions (3% O2) for 6 h followed by reoxygenation at room air (21% O2) for 1 h, but in the presence or absence of PDTC (10, 50, 100�M). EMSA for NF�B was performed as described above. A normoxic control is shown in lane 1 of this representative assay.


tin or ICAM-1 expression in HMEC-1. IgG controlsfor these studies revealed no significant fluorescence(data not shown).

DISCUSSION

Abrupt return of circulation to vascular beds follow-ing periods of ischemia or intense hypoxia commonlyresults in significant vascular injury and the frequentdevelopment of organ failure [1–4]. Examination of tis-sues obtained from organ beds following onset of reper-fusion injury has documented biological events consis-tent with an acute inflammatory response (i.e., loss ofvascular endothelial cell integrity, parenchymal edema,and intensive neutrophil sequestration within microvas-culature) [26–28]. Granger and colleagues and othershave suggested that reperfusion injury at the level of themicrocirculation arises from the generation of ROSformed within tissue following reintroduction of oxygenafter ischemic events [29,30]. Multiple studies now re-veal that restoration of molecular oxygen to physiologiclevels following periods of hypoxia results in generationof ROS in many critical cell types [31–33]. However, thecontribution of microvascular endothelium to the gener-ation of ROS and the biological consequences of thatproduction are only now becoming defined.

In the current studies, we examined molecular eventsoccurring in microvascular endothelium during periodsof posthypoxic or reoxygenation stress. The endothelialcells used for these experiments are an established hu-man dermal microvascular cell line (HMEC-1) first re-ported by Ades and colleagues [18]. Numerous publishedreports now show that HMEC-1 exhibit biochemical andmolecular characteristics comparable to primary humanendothelial cell lines [34–37]. Our model system incor-porated a 6 h hypoxic period (3% oxygen) followed byan abrupt return to atmospheric oxygen concentration(21%), plus 5% carbon dioxide and a balance of nitro-gen. HMEC-1 exposed to these conditions exhibitedimmediate and significant generation of oxidant stress asassessed by H2DCF-DA epifluorescence and spectroflu-

ometry (Fig. 1B and E). Within this time frame followingonset of reoxygenation (60 s), HMEC-1 showed intenseperi-nuclear mitochondrial epifluorescence (Fig. 1D).The location of such intense perinuclear epifluorescencesuggests that mitochondria may serve as a significantsource of ROS generated in HMEC-1 under these con-ditions [38,39]. When reoxygenating HMEC-1 were ex-posed to the antioxidant pyrrolidine dithiocarbamate,H2DCF-DA epifluorescence was abolished. Little or noepifluorescence was observed in normoxic HMEC-1controls. These studies suggest that HMEC-1 are sub-jected to significant oxidant stress under the modelingconditions established.

A key finding in these studies was the rapid andsignificant activation of NF-�B in cells subjected toreoxygenation stress. Electrophoretic mobility shift stud-ies (EMSA) revealed that within 60 min, reoxygenatingcells exhibited a greater than 5-fold increase of NF-�Bprotein within nuclear extracts when compared to nor-moxic controls (p � .001, Fig. 2A). Very importantly,NF-�B activation in our model system was a dynamicprocess. Beyond 60 min, NF-�B activation in reoxygen-ating cells diminished rapidly, returning to control levelsby 120 min. Once downregulation of NF�B had oc-curred, it remained unchanged for an additional 10 h ofobservation (Fig. 2A). An identical experimental ap-proach revealed that AP-1 exhibited a similar pattern ofactivation, however, peak activation was delayed in timeby 60 min. AP-1 DNA binding, as detected by EMSA, isa measure of the quantity of AP-1 present in the nucleus.Increases in AP-1 DNA binding equate to transcriptionof fresh AP-1 subunits [40]. Unlike AP-1, a reservoir ofpresynthesized NF�B protein exists in cell cytoplasmheld in an inactivated state until signaling events occur.As observed in these studies, NF�B was rapidly acti-vated. De-novo synthesis of AP-1 subunits resulted indelayed peak AP-1 DNA binding. Similar to NF�B,AP-1 activation rapidly returned to levels at or belowcontrols in the ensuing 10 h. The abrupt appearance ofNF-�B and AP-1 protein in the nuclei of reoxygenatingHMEC-1 followed by their equally rapid disappearance

Fig. 3. Detection of I�B� during reoxygenation. HMEC-1 cells were grown to confluence in 35 mm dishes. They were treated withmedia alone (*) or with TNF-� (10 ng/ml) for 15, 45, and 90 min. Alternately, they were maintained under normoxic conditions (#)or exposed to hypoxic conditions (3% O2) for 6 h followed by reoxygenation at room air (21% O2) for 15, 45, and 90 min. Cytosolicextracts (10 �g protein) were separated by SDS-Polyacrylamide gel electrophoresis (10%) and electrophoretically transferred topolyvinylidene fluoride membranes (Immobilon P, 0.45 mm pore size). Immunodetection was performed using a primary I�B�antibody and the Renaissance Western Blot Chemiluminescence Reagent Plus.


suggests the presence of a highly regulated process fortranscription factor movement, perhaps a shuttling mech-

anism which becomes activated once cells are subjectedto the level of oxidative stress produced by our model.To gain further insight into the mechanism of NF-�Bactivation in reoxygenating HMEC-1, both reoxygenat-ing and normoxic cells were exposed to the thiol antiox-idant pyrrolidine dithiocarbamate (PDTC). PDTC iswidely employed as an inhibitor of NF-�B and may havetherapeutic potential in inflammatory diseases linked toredox imbalance and generation of ROS [41,42]. Themechanism of action of PDTC is conferred by its anti-oxidant properties through a concentration-dependent in-crease in cellular glutathione (GSH) levels. PDTC-in-duced increases in GSH occur through the activation ofgamma-glutamylcysteine synthetase [43]. These studiesreveal that PDTC produced significant concentration-

Fig. 4. Reoxygenation stress promotes significant increases in neutro-phil adhesion. HMEC-1 were grown to confluence in 48 well cultureplates and then subjected to 6 h of hypoxia. Medium overlying bothhypoxic HMEC-1 and normoxic HMEC-1 cultures was removed andreplaced with fresh growth medium. Cultures were then incubated at37°C, in 21% oxygen, 5% CO2, balance N2 for 120 min. At 120 minCalcein AM-labeled PMN were added to all wells and allowed toadhere for 10, 20, 30, and 45 min. Wells were washed three times andlabeled adherent PMN remaining in each well imaged and enumeratedat 10� magnification using fluorescence microscopy as described inMaterials and Methods. For each condition the numbers of adherentcells remaining in triplicate wells is reported. The results are expressedas cells per high-powered field. At each adherence time examined,reoxygenating HMEC-1 promoted significantly greater PMN adhesionthan did normoxic controls (p � .05).

Fig. 5. HMEC-1 were grown to confluence in 48 well culture plates andthen subjected to hypoxia (6 h, 3% O2)/reoxygenation (120 min, 21% O2)but in the presence of 10, 20, 30, 40, and 50 �M pyrrolidine dithiocar-bamate (PDTC). Calcein AM-labeled PMN were added to all wells andallowed to adhere for 20 min. Wells were washed three times and labeledadherent PMN remaining in each well imaged and enumerated at 10�magnification using fluorescence microscopy as described in Materials andMethods. The results are expressed as percent normoxic controls � SEMof three independent observations (*p � .001).

Fig. 6. IL-8 secretion by HMEC-1 measured under conditions orreoxygenation and cytokine stress. (A) HMEC-1 were exposed tohypoxia (3% O2) followed by return to 21% oxygen for a period of upto 8 h. Culture medium was then analyzed for IL-8 by ELISA. (A)Studies show that IL-8 secretion by posthypoxic HMEC-1 are notdifferent from normoxic controls (n � 5). (B) HMEC-1 were exposedto medium containing increasing concentrations of TNF� (0.01–100ng/ml) or medium alone for 4 h. Medium was then analyzed for IL-8by ELISA. These studies show that TNF�-exposed HMEC-1 secretedsignificantly greater IL-8 (*p � .001 vs. media control) at everyconcentration tested.


dependent attenuation of NF-�B activation in reoxygen-ating endothelium (Fig. 2B). At the highest concentrationof PDTC employed in these experiments (100 �M) a90 � 2% overall reduction of NF-�B activation wasobserved.

Of major importance in these studies was the findingthat NF-�B activation induced by ROS generation wasnot associated with cytoplasmic degradation of I�B�. Inthe current study and in prior publications, we and othershave shown that activation of endothelium via cytokineexposure (e.g., TNF-�, interleukin-1�) results in degra-dation of I�B� via a proteosome-mediated mechanism[25,44]. Western blot analysis of cytoplasmic extractsfrom reoxygenating HMEC-1 in the current study re-vealed no changes in I�B� protein levels during theperiod of NF-�B activation (Fig. 3). Canty and col-leagues examined both human myocardium and umbili-cal vein endothelium subjected to reoxygenation stressand found that NF-�B activation occurred without I�B�degradation [45]. Zwacka et al. examined livers ofBALB/c mice subjected to ischemia/reperfusion andshowed substantial NF�B activation that was indepen-dent of proteolytic cleavage of I�B� [46]. Unpublished

studies from the authors’ laboratory demonstrate thattyrosine phosphorylation of I�B� represents a proteoly-sis-independent mechanism of NF�B activation that maybe targeted for prevention of reoxygenation-mediatedinjury, a process which may occur without impactingnormal inflammatory responses. Thus, the data from thecurrent study suggests a clear stimulus specificity withrespect to the fate of IkB� protein and the activation ofNF�B.

An important goal for these studies was to correlatemolecular events occurring within our HMEC-1 modelsystem with neutrophil adhesion, an endothelial cellfunction necessary for the genesis of acute inflammation.For these studies, identical numbers of Calcein-AM-labeled neutrophils were incubated with cultures ofreoxygenating and normoxic HMEC-1. Our results showthat reoxygenation stress induced a 500% increase inPMN adhesion to HMEC-1 at 120 min compared tonormoxic controls (Fig. 4). Kokura and colleagues, uti-lizing a similar model of hypoxia/reoxygenation, re-ported that human umbilical vein endothelium exhibitedmarked depletion of GSH during a time of enhancedPMN adhesion, suggesting a key role for oxidant stress

Fig. 7. E-selectin and ICAM-1 expression in normoxic and posthypoxic HMEC-1. (A) HMEC-1 cells exposed to normoxic conditionsimmunostained for E-selectin. (B) HMEC-1 cells exposed to hypoxic conditions (3% O2) for 6 h followed by re-oxygenation at roomair (21% O2) for 2 h and immunostained for E-selectin. (C) HMEC-1 cells exposed to normoxic conditions and immunostained forICAM-1. (D) HMEC-1 cells exposed to hypoxic conditions (3% O2) for 6 h followed by re-oxygenation at room air (21% O2) for 2 hand immunostained for ICAM-1 (60� magnification).


in the genesis of PMN adhesion [47]. When oxidantstressed HMEC-1 were exposed to the thiol antioxidantPDTC, PMN adhesion diminished dramatically, return-ing to those levels observed in normoxic controlHMEC-1 (Fig. 5). Thus, neutrophil adhesion in thismodel system was clearly driven by ROS generatedduring the reoxygenation phenomenon. Immunofluores-cence studies performed on HMEC-1 under identicalconditions of reoxygenation stress reveal significant up-regulation of both E-selectin and ICAM-1 compared tonormoxic controls (Fig. 7). These results indicate thatendothelial cell molecules necessary for both rolling andtethering adhesion are upregulated by the rexoygenationstress delivered in this model system. The results re-ported here are similar to those found by Kokura andcolleagues, who suggested that thiol imbalance plays acritical role in expression of adhesion molecules in en-dothelium exposed to Anoxia/reoxygenation [48].

Significant evidence now shows that IL-8 plays anessential, if not critical, role in promoting and maintain-ing inflammation [49,50]. Rainger and colleagues havereported that IL-8 secretion by endothelium exposed toreoxygenation stress is essential for stimulating the im-mobilization and tethering of PMN in a flow-simulatedenvironment [12,51]. However, the findings reportedhere differ somewhat from prior studies. HMEC-1 ex-posed to the oxidant stress conditions described secretedvery little IL-8—quantities that were not different fromnormoxic controls (Fig. 6A). These results could not beexplained by any changes in cell mortality. Careful anal-ysis of cell viability in our model system revealed 99%viability in both reoxygenated and normoxic cell popu-lations. In contradistinction, HMEC-1 activated by oxi-dant stress induced by TNF�, exhibited significant (p �.001), concentration-dependent, increases in IL-8 protein(Fig. 6B). An explanation for our results may be found inthe activation kinetics of transcription factors NF�B andAP-1 reported in Fig. 2A. Our results would suggest thatthe signaling pathways activated in HMEC-1 by ROSgeneration as reported here fail to achieve critical levelsof simultaneous NF�B and AP-1 activation. While thispattern of activation is sufficient for enhanced PMNadhesion in reoxygenating HMEC-1 (Fig. 4), it fails topromote IL-8 secretion. Recently, we reported that crit-ical interactions between transcription factors NF�B,AP-1, and NF-IL-6 were necessary for endothelial IL-8secretion to occur [52]. Lakshminarayanan and col-leagues found identical results with respect to IL-8 se-cretion in HMEC-1 following exposure to exogenouslysupplied oxidative stress (H2O2) [53].

In conclusion, this study has sharply focused on mo-lecular events occurring in microvascular endotheliumfollowing onset of reoxygenation-associated oxidativestress. These conditions were produced simply by return-

ing hypoxic cells to atmospheric pressures (760 � 5 mmHg average barometric pressure in Richmond, VA). Inno experiments were cells exposed to added or supple-mental oxygen. As described above, the conditions es-tablished in this model system resulted in extremes ofoxygenation, genuinely providing an environment ofsubstantially changed partial pressures of oxygen.Reoxygenation stress as modeled here induces signifi-cant increases in microvascular endothelial oxidative ac-tivity. This activity rapidly modifies quiescent HMEC-1inducing nonsimultaneous NF-�B and AP-1 activation.These events, while sufficient to significantly upregulateneutrophil adhesion, fail to promote IL-8 secretion abovenormoxic control cells. These results strongly suggestalternative oxidative signaling pathways produced byreoxygenation and cytokine exposure.

Acknowledgements — This research was supported by funds from theNational Institutes of Health (HL-61359 and HL-10355) and the Vir-ginia Thoracic Society.

REFERENCES

[1] Roumen, R. M.; Hendriks, T.; van der Ven-Jongekrijg, J.; Nieu-wenhuijzen, G. A.; Sauerwein, R. W.; van der Meer, J. W.; Goris,R. J. Cytokine patterns in patients after major vascular surgery,hemorrhagic shock, and severe blunt trauma. Relation with sub-sequent adult respiratory distress syndrome and multiple organfailure. Ann. Surg. 218:769–776; 1993.

[2] Welborn, M. B.; Oldenburg, H. S.; Hess, P. J.; Huber, T. S.;Martin, T. D.; Rauwerda, J. A.; Wesdorp, R. I.; Espat, N. J.;Copeland, E. M. 3rd; Moldawer, L. L.; Seeger, J. M. The rela-tionship between visceral ischemia, proinflammatory cytokines,and organ injury in patients undergoing thoracoabdominal aorticaneurysm repair. Crit. Care. Med. 28:3191–3197; 2000.

[3] Gavin, J. B.; Maxwell, L.; Edgar, S. G. Microvascular involve-ment in cardiac pathology. J. Mol. Cell. Cardiol. 30:2531–2540;1998.

[4] Jean, W. C.; Spellman, S. R.; Nussbaum, E. S.; Low, W. C.Reperfusion injury after focal cerebral ischemia: the role of in-flammation and the therapeutic horizon. Neurosurgery 43:1382–1396; discussion 1396–1397; 1998.

[5] Thiagarajan, R. R.; Winn, R. K.; Harlan, J. M. The role ofleukocyte and endothelial adhesion molecules in ischemia-reper-fusion injury. Thromb. Haemost. 78:310–314; 1997.

[6] Zimmerman, B. J.; Grisham, M. B.; Granger, D. N. Mechanismsof oxidant-mediated microvascular injury following reperfusionof the ischemic intestine. Basic Life Sci. 49:881–886; 1988.

[7] Zimmerman, B. J.; Granger, D. N. Mechanisms of reperfusioninjury. Am. J. Med. Sci. 307:284–292; 1994.

[8] Menger, M. D.; Pelikan, S.; Steiner, D.; Messmer, K. Microvas-cular ischemia-reperfusion injury in striated muscle: significanceof “reflow paradox.” Am. J. Physiol. 263:H1901–H1906; 1992.

[9] Steinbauer, M.; Harris, A. G.; Leiderer, R.; Abels, C.; Messmer,K. Impact of dextran on microvascular disturbances and tissueinjury following ischemia/reperfusion in striated muscle. Shock9:345–351; 1998.

[10] Topham, M. K.; Carveth, H. J.; McIntyre, T. M.; Prescott, S. M.;Zimmerman, G. A. Human endothelial cells regulate polymor-phonuclear leukocyte degranulation. FASEB J. 12:733–746;1998.

[11] Oz, M. C.; Liao, H.; Naka, Y.; Seldomridge, A.; Becker, D. N.;Michler, R. E.; Smith, C. R.; Rose, E. A.; Stern, D. M.; Pinsky,D. J. Ischemia-induced interleukin-8 release after human hearttransplantation. A potential role for endothelial cells. Circulation92(Suppl. 9):II428–II432; 1995.


[12] Rainger, G. E.; Fisher, A.; Shearman, C.; Nash, G. B. Adhesion offlowing neutrophils to cultured endothelial cells after hypoxia andreoxygenation in vitro. Am. J. Physiol. 269:H1398–H1406; 1995.

[13] Bertuglia, S.; Colantuoni, A.; Intaglietta, M. Effect of leukocyteadhesion and microvascular permeability on capillary perfusionduring ischemia-reperfusion injury in hamster cheek pouch. Int. J.Microcirc. Clin. Exp. 13:13–26; 1993.

[14] Muller, J. M.; Rupec, R. A.; Baeuerle, P. A. Study of regulationby NF-kappa B and AP-1 in response to reactive oxygen inter-mediates. Methods 11:301–312; 1997.

[15] Dalton, T. P.; Shertzer, H. G.; Puga, A. Regulation of geneexpression by reactive oxygen. Annu. Rev. Pharmacol. Toxicol.39:67–101; 1999.

[16] Baeuerle, P. A.; Henkel, T. Function and activation of NF-kappaB in the immune system. Annu. Rev. Immunol. 12:141–179; 1994.

[17] Ziegler, D. M. Role of reversible oxidation-reduction of enzymethiols-disulfides in metabolic regulation. Annu. Rev. Biochem.54:305–329; 1985.

[18] Ades, E. W.; Candal, F. J.; Swerlick, R. A.; George, V. G.;Summers, S.; Bosse, D. C.; Lawley, T. J. HMEC-1: establishmentof an immortalized human microvascular endothelial cell line.J. Invest. Dermatol. 99:683–690; 1992.

[19] Papadopoulos, N. G.; Dedoussis, G. V.; Spanakos, G.; Gritzapis,A. D.; Baxevanis, C. N.; Papamichail, M. An improved fluores-cence assay for the determination of lymphocyte-mediated cyto-toxicity using flow cytometry. J. Immunol. Methods 177:101–111; 1994.

[20] Hempel, S. L.; Buettner, G. R.; O’Malley, Y. Q.; Wessels, D. A.;Flaherty, D. M. Dihydrofluorescein diacetate is superior for de-tecting intracellular oxidants: comparison with 2�,7�-dichlorodi-hydrofluorescein diacetate, 5(and 6)-carboxy-2�,7�-dichlorodihy-drofluorescein diacetate, and dihydrorhodamine 123. Free Radic.Biol. Med. 27:146–159; 1999.

[21] Fowler, A. A.; Fisher, B. J.; Sweeney, L. B.; Wallace, T. J.;Natarajan, R.; Ghosh, S. S.; Ghosh, S. Nitric oxide regulates IL-8gene expression in activated endothelium by inhibiting NF-kBbinding to DNA: effects on endothelial function. Biochem. CellBiol. 77:201–208; 1999.

[22] Dignam, J. K.; Lebovitz, R. M.; Roeder, R. G. Accurate tran-scription initiation by RNA polymerase II in a soluble extractfrom isolated mammalian nuclei. Nucleic Acids Res. 11:1475–1489; 1983.

[23] Harant, H.; De Martin, R.; Andrew, P. J.; Foglar, E.; Dittrich, C.;Lindley, I. J. D. Synergistic activation of interleukin-8 genetranscription by all-trans-retinoic acid and tumor necrosis factor-ainvolves the transcription of factor NF-kB. J. Biol. Chem. 271:26954–26961; 1996.

[24] Braut-Boucher, F.; Pichon, J.; Rat, P.; Adolphe, M.; Aubery, M.;Font, J. A non-isotopic, highly sensitive, fluorimetric, cell-celladhesion microplate assay using calcein AM-labeled lympho-cytes. J. Immunol. Methods 178:41–51; 1995.

[25] Fowler, A. A.; Fisher, B. J.; Centor, R. M.; Carchman, R. A.Development of the adult respiratory distress syndrome: progres-sive alteration of neutrophil chemotactic and secretory processes.Am. J. Pathol. 116:427–435; 1984.

[26] Koksoy, C.; Kuzu, M. A.; Kuzu, I.; Ergun, H.; Gurhan, I. Role oftumour necrosis factor in lung injury caused by intestinal isch-aemia-reperfusion. Br. J. Surg. 88:464–468; 2001.

[27] Colletti, L. M.; Cortis, A.; Lukacs, N.; Kunkel, S. L.; Green, M.;Strieter, R. M. Tumor necrosis factor up-regulates intercellularadhesion molecule 1, which is important in the neutrophil-depen-dent lung and liver injury associated with hepatic ischemia andreperfusion in the rat. Shock 10:182–191; 1998.

[28] Childs, E. W.; Wood, J. G.; Smalley, D. M.; Hunter, F. A.;Cheung, L. Y. Leukocyte adherence and sequestration followinghemorrhagic shock and total ischemia in rats. Shock 11:248–252;1999.

[29] Granger, D. N. Mechanisms of microvascular injury. In: Johnson,L. R.; Alpers, D. H.; Christensen, J.; Jacobson, E. D.; Walsh,J. H., eds. Physiology of the gastrointestinal track, 3rd ed. NewYork: Lippincott-Raven; 1994:1693–1722.

[30] Cuzzocrea, S.; Riley, D. P.; Caputi, A. P.; Salvemini, D. Antiox-idant therapy: a new pharmacological approach in shock, inflam-mation, and ischemia/reperfusion injury. Pharmacol. Rev. 53:135–159; 2001.

[31] Mohazzab-H, K. M.; Kaminski, P. M.; Wolin, M. S. Lactate andPO2 modulate superoxide anion production in bovine cardiacmyocytes: potential role of NADH oxidase. Circulation 96:614–620; 1997.

[32] Lounsbury, K. M.; Hu, Q.; Ziegelstein, R. C. Calcium signalingand oxidant stress in the vasculature. Free Radic. Biol. Med.28:1362–1369; 2000.

[33] Mason, R. B.; Pluta, R. M.; Walbridge, S.; Wink, D. A.; Oldfield,E. H.; Boock, R. J. Production of reactive oxygen species afterreperfusion in vitro and in vivo: protective effect of nitric oxide.J. Neurosurg. 93:99–107; 2000.

[34] Goebeler, M.; Yoshimura, T.; Toksoy, A.; Ritter, U.; Brocker,E. B.; Gillitzer, R. The chemokine repertoire of human dermalmicrovascular endothelial cells and its regulation by inflammatorycytokines. J. Invest. Dermatol. 108:445–451; 1997.

[35] Nagai, Y.; Tasaki, H.; Takatsu, H.; Nihei, S.; Yamashita, K.;Toyokawa, T.; Nakashima, Y. Homocysteine inhibits angiogene-sis in vitro and in vivo. Biochem. Biophys. Res. Commun. 281:726–731; 2001.

[36] Kirton, C. M.; Nash, G. B. Activated platelets adherent to anintact endothelial cell monolayer bind flowing neutrophils andenable them to transfer to the endothelial surface. J. Lab. Clin.Med. 136:303–313; 2000.

[37] Willam, C.; Koehne, P.; Jurgensen, J. S.; Grafe, M.; Wagner,K. D.; Bachmann, S.; Frei, U.; Eckardt, K. U. Tie2 receptorexpression is stimulated by hypoxia and proinflammatory cyto-kines in human endothelial cells. Circ. Res. 87:370–377; 2000.

[38] Esposti, M. D.; Hatzinisiriou, I.; McLennan, H.; Ralph, S. Bcl-2and mitochondrial oxygen radicals. New approaches with reactiveoxygen species-sensitive probes. J. Biol. Chem. 274:29831–29837; 1999.

[39] Soultanakis, R. P.; Melamede, R. J.; Bespalov, I. A.; Wallace,S. S.; Beckman, K. B.; Ames, B. N.; Taatjes, D. J.; Janssen-Heininger, Y. M. Fluorescence detection of 8-oxoguanine innuclear and mitochondrial DNA of cultured cells using a recom-binant Fab and confocal scanning laser microscopy. Free Radic.Biol. Med. 28:987–998; 2000.

[40] Angel, P.; Karin, M. The role of Jun, Fos and the AP-1 complexin cell-proliferation and transformation. Biochim. Biophys. Acta1072:129–157; 1991.

[41] Sasaki, H.; Zhu, L.; Fukuda, S.; Maulik, N. Inhibition of NFkappa B activation by pyrrolidine dithiocarbamate prevents invivo hypoxia/reoxygenation-mediated myocardial angiogenesis.Int. J. Tissue React. 22:93–100; 2000.

[42] Natarajan, R.; Ghosh, S.; Fisher, B. J.; Diegelmann, R. F.; Willey,A.; Walsh, S.; Graham, M. F.; Fowler, A. A. 3rd. Redox imbal-ance in crohn’s disease intestinal smooth muscle cells causesNF-kappa B-mediated spontaneous interleukin-8 secretion. J. In-terferon Cytokine Res. 21:349–359; 2001.

[43] Wild, A. C.; Mulcahy, R. T. Pyrrolidine dithiocarbamate up-regulates the expression of the genes encoding the catalytic andregulatory subunits of gamma-glutamylcysteine synthetase andincreases intracellular glutathione levels. Biochem. J. 338:659–665; 1999.

[44] Parry, G. C.; Martin, T.; Felts, K. A.; Cobb, R. R. IL-1beta-induced monocyte chemoattractant protein-1 gene expression inendothelial cells is blocked by proteasome inhibitors. Arterio-scler. Thromb. Vasc. Biol. 18:934–940; 1998.

[45] Canty, T. G.; Boyle, E. M.; Farr, A.; Morgan, E. N.; Verrier,E. D.; Pohlman, T. H. Oxidative stress induces NF-kappaB nu-clear translocation without degradation of IkappaBalpha. Circu-lation 100:361–364; 1999.

[46] Zwacka, R. M.; Zhang, Y.; Zhou, W.; Halldorson, J.; Engelhardt,J. F. Ischemia/reperfusion injury in the liver of BALB/c miceactivates AP-1 and nuclear factor kappaB independently of Ikap-paB degradation. Hepatology 28:1022–1030; 1998.

[47] Kokura, S.; Wolf, R. E.; Yoshikawa, T.; Granger, D. N.; Aw,


T. Y. Molecular mechanisms of neutrophil-endothelial cell adhe-sion induced by redox imbalance. Circ. Res. 84:516–524; 1999.

[48] Kokura, S.; Rhoads, C. A.; Wolf, R. E.; Yoshikawa, T.; Granger,D. N.; Aw, T. Y. NF kappa B signaling in posthypoxic endothelialcells: relevance to E-selectin expression and neutrophil adhesion.J. Vasc. Res. 38:47–58; 2001.

[49] Mukaida, N. Interleukin-8: an expanding universe beyond neu-trophil chemotaxis and activation. Int. J. Hematol. 72:391–398;2000.

[50] Atta-ur-Rahman; Harvey, K.; Siddiqui, R. A. Interleukin-8: anautocrine inflammatory mediator. Curr. Pharm. Des. 5:241–253;1999.

[51] Rainger, G. E.; Fisher, A. C.; Nash, G. B. Endothelial-borneplatelet-activating factor and interleukin-8 rapidly immobilizerolling neutrophils. Am. J. Physiol. 272:H114–H122; 1997.

[52] Natarajan, R.; Gupta, S.; Fisher, B. J.; Ghosh, S.; Fowler, A. A.Nitric oxide suppresses IL-8 transcription by inhibiting C-JunN-Terminal Kinase-Induced AP-1 activation. Exp. Cell. Res. 266:203–212; 2001.

[53] Lakshminarayanan, V.; Beno, D. W. A.; Costa, R. H.; Roebuck,K. A. Differential regulation of interleukin-8 and intercellularadhesion molecule-1 by H2O2 and tumor necrosis factor-� inendothelial and epithelial cells. J. Biol. Chem. 272:32910–32918;1997.

ABBREVIATIONS

AP-1—Activator protein-1HMEC-1—Human microvascular endothelial cell-1I�B�—Inhibitor kappa B alphaIL-8—Interleukin-8NF�B—Nuclear factor kappa BPDTC—Pyrrolidine dithiocarbamatePMN—Polymorphonuclear neutrophilROS—Reactive oxygen species


reoxygenating microvascular endothelium exhibits temporal dissociation of nf-κb and ap-1 activation

Documents