Hydrogen peroxide and superoxide modulate leukocyte adhesion molecule expression and leukocyte endothelial adhesion

BB

ELSEVIER

Biochi~ic~a et BiophysicaAEta Biochimica et Biophysica Acta 1310 (1996) 251-259

Hydrogen peroxide and superoxide modulate leukocyte adhesion molecule expression and leukocyte endothelial adhesion Aureliano Fraticelh a, Carlos V. Serrano Jr. a, Bruce S. Bochner u Maurizio C. Capogrossi c Jay L. Zweier a., a Molecular and Cellular Biophysics Laboratories and EPR Center, Department of Medicine, Division of Cardiology, The Johns Hopkins Medical Institutions, Johns Hopkins Bayview Medical Center, Baltimore, MD 21224, USA b Division of Clinical Immunology, Department of Medicine, The Johns Hopkins Medical Institutions, Johns Hopkins Bayview Medical Center, Baltimore, MD 21224, USA c Laboratory of Cardiovascular Science, Gerontology Research Center, National Institute on Aging, Baltimore, MD 21224, USA Received 3 August 1995; accepted 17 October 1995

Abstract While endothelial oxidant generation and subsequent leukocyte chemotaxis and activation are important mechanisms of tissue damage in ischemic organs, it is not known if oxidant generation may be involved in triggering the subsequent leukocyte-mediated injury which occurs. Questions remain w]hether particular oxidants and oxygen-free radicals are capable of modulating the expression of leukocyte adhesion molecules and effecting leukocyte endothelial adhesion. Studies were performed to determine the effect of different biologically occurring oxidant molecules and oxygen free radicals including: .02, .OH, and H202 on the expression of integrin and selectin adhesion molecules on the surface of human PMNs and to determine the effect of these alterations on PMN adhesion to the endothelium. Adhesion molecule expression on the surface of human PMNs was measured by immunofluorescence flow cytometry. Electron paramagnetic resonance spectroscopy was applied to characterize the presence of exogenous free radical generation as well as that from activated PMNs. It was observed that these oxidants can cause up-regulation of CD1 lb and CDI8 expression with shedding of L-selectin. The kinetics and dose-response of these effects were analyzed and their functional significance determined by measuring PMN adhesion to cultured human aortic endothelial monolayers. These studies demonstrate that oxygen free radicals and non-radical oxidants can directly trigger PMN activation and adhesion to vascular endothelium. Keywords: Free radical; Oxidant; Inflammation; Selectin expression; Integrin expression

I. Introduction Oxygen free radicals as well as non-radical oxidants have been shown to be important mediators of cellular injury in a number of pathophysiological processes including tissue damage during ischemia and reperfusion, the pathogenesis of atherosclerosis, and inflammatory injury [1-6]. In the setting of postischemic injury endothelial oxidant generation is followed by recruitment and infiltration of leukocytes into the effected tissue. While it has been hypothesized that these radical or non-radical oxidants may be involved in triggering the subsequent inflammatory responses which occur with the infiltration of polymorphonuclear leukocytes, PMNs, the basis and

* Corresponding author. Fax: + 1 (410) 550-2448. 0167-4889/96/$15.00 © 1996, Elsevier Science B.V. All rights reserved SSDI 0 1 6 7 - 4 8 8 9 ( 9 5 ) 0 0 1 6 9 - 7

molecular mechanism by which these two sequential processes occur, is unknown. The process of PMN infiltration within tissues is initiated by adhesion to the endothelium followed by endothelial transmigration [7]. Thus, adhesion of the PMN to the endothelial surface is an essential first step in the mechanism of all inflammatory processes. Leukocyte-endothelial cell adhesion involves two major classes of adhesion molecules, the selectins and the integrins. L-selectin is a carbohydrate binding adhesion molecule protein which is constitutively expressed on the surface of non-activated granulocytes and shed after activation. It is thought to be involved in the process of PMN rolling on the endothelial surface under conditions of shear stress [8-10]. The /32-integrins consist of a family of heterodimers, each of which has a common /32-subunit designated CD18, and one of three known c~-subunits designated C D l l a (LFA-1), C D l l b (Mac-l), and C D l l c

252

A. Fraticelli et al. / Biochimica et Biophysica Acta 1310 (1996) 251-259

[p150,95] [7,11]. The surface expression of Mac-I and p150,95 are up-regulated when PMNs are activated [11]. It has been shown that firm PMN adhesion is mediated by /32-integrins. The Mac-1 hetero-dimer is thought to be the most functionally important integrin and an endothelial counter-ligand for Mac-1 is ICAM-1, intercellular adhesion molecule-1 [12,13]. It has been demonstrated that oxygen-derived free radicals and other non-paramagnetic oxidants including: superoxide, . 0 2 ; hydroxyl radicals, •OH; hydrogen peroxide, H202; and hypochlorous anion, HOCI-, are generated in reoxygenated vascular endothelial cells or in activated PMNs with sufficient concentrations to cause cellular injury and death [14-16]. This suggests that the endothelium may be a particularly important site of oxidant generation [14]. A link between the generation of oxidant species and leukocyte recruitment to the reperfusion site has been suggested from the observation that both superoxide dismutase, SOD, and the xanthine oxidase inhibitor, allopurinol, reduce PMN infiltration in ischemic-reperfused tissues [17]. By intravital microscopy it has been demonstrated that the reperfusion-induced leukocyte adherence to the endothelial wall is decreased by various antioxidative enzymes or compounds that either scavenge or prevent the formation of free radicals and other oxidants [18,19]. The goal of this study was to determine the effect of different biologically occurring oxidant molecules and oxygen free radicals including: . 0 2, .OH and H202 on the expression of integrin and selectin adhesion molecules on the surface of PMNs and to determine the functional effect of these alterations on PMN adhesion to the endothelium. These studies demonstrate that free radicals and oxidants can directly trigger PMN activation and adhesion to vascular endothelium.

2. Materials and methods

2.1. Leukocyte purification and preparation Human PMNs were isolated from the blood of healthy subjects of either sex by Percoll density gradient centrifugation [20]. Freshly sampled venous blood (10 ml) was added to 25 ml of 0.9% NaC1 and 0.8 ml of 0.1 M EDTA (ethylene diamine tetraacetic acid). The diluted blood was layered on 9 ml of Percoll-containing solution, with a specific density of 1.08 g / m l at 20°C. This solution consists of Percoll (56% volume) diluted with a 10 times concentrated PIPES (piperazine-N, N'-bis-[2-ethanosulfonic-acid]) buffer. PIPES buffer contains 25 mM PIPES, 110 mM NaCI, 5 mM KC1 (pH 7.4). After room temperature centrifugation at 400 × g for 20 rain, supernatant was discarded. All subsequent steps were performed at 4°C. The remaining red pellet underwent hypotonic lysis, by the addition of 18 ml HzO. After 30 s, osmolarity was restored with 2 ml of 10 times concentrated PIPES

buffer, followed by centrifugation at 600 × g for 5 min. Supernatant was discarded, and the residual pellet underwent a second cycle of hypotonic lysis and centrifugation, followed by two washes in 15 ml of PAG buffer (PIPES buffer with 0.003% human serum albumin and 0.1% dextrose) and centrifugation at 400 × g for 6 min. The final pellet was then suspended in 1 ml of PAG-CM buffer (PAG buffer with 1 mM CaC12 and 1 mM MgCI2). This procedure yielded 15-35 × 10 6 PMNs for each 10 ml of blood. Differential count after May Grumwald-Giemsa staining confirmed that over 95% of the cells were PMNs; with the remaining predominantly eosinophils. Viability of the isolated PMNs, estimated by erythrosin B exclusion, was greater than 99%.

2.2. Exposure of PMNs to oxidants PMNs suspended in PAG-CM buffer were challenged with oxidant species including H 2 0 2 , .OH, .O 2, or HOC1- at 37°C in a shaking water-bath. The tubes were then centrifuged, and the pellets washed twice with ice-cold Dulbecco's phosphate-buffered saline containing bovine serum albumin (138 mM NaC1, 27 mM KC1, 11 mM K H z P O 4, 8 mM Na2HPO 4, and 0.2% BSA; pH 7.4; PBS + BSA). The expression of the PMN adhesion molecules was then measured. The effect of H202 on the PMN adhesion molecules CD1 lb, CD18, and L-selectin was studied in dose-response and time course experiments. Concentrations of 0.1-10.0 mM H 2 0 2 were studied with an incubation period of 30 min. For time course experiments, 1, 5, 15, 30, and 60 rain time points were used, with a H202 concentration of 5.0 mM. In order to rapidly terminate H202 stimulation, to block cellular processes, and fix cell membranes, aliquots of the PMN suspension were drawn and immediately placed in ice-cold PAG buffer containing catalase (500 U/ml), EDTA (0.1 M) and 0.25% paraformaldehyde. After 2.5--3 min, PMNs were washed with PBS + BSA, prior to antibody labeling. As previously reported, •OH can be formed from H 2 0 2 in the presence of iron chelates which facilitate the Fenton reaction, as shown in the following reactions [21]: H 2 0 2 -1- 2Fe3+--* 2FEZ+ + O 2 --b 2H +

2(Fe 2+ + H202 ~ Fe 3+ + OH- + • OH) ]E 3H202 --* 2 • OH + 02 + 2H20. Hydroxyl radical generation was triggered by addition of the iron chelate Fe3+-nitrilotriacetate(l:2), 20 /zM, to the PMN suspension prior to the H 2 0 z (5.0 mM). Superoxide was generated by the reaction of xanthine oxidase (40 m U / m l ) with its substrate xanthine. To prevent the iron-catalyzed production of •OH from H 2 0 z, the reaction was performed in the presence of catalase (500 U / m l ) and deferoxamine (1.0 mM) [22]. Hypochlorous anion (1.0 mM) was generated by adding

A. FraticeUi et al. / Biochimica et Biophysica Acta 1310 (1996) 251-259

sodium hypochlorite (NaOC1, 1.0 raM) to the PMN suspension. Taurine (5.0 mM), a HOC1- scavenger, or sodium azide (1.0 mM), an inhibitor of myeloperoxidase, were also used to characterize the H20 z and the HOC1--induced changes on PMN adhesion molecule expression [23,24]. 2.3. Measurement of adhesion molecule expression

Indirect immunofluorescence and flow-cytometry was performed on aliquots of 10 6 PMNs, suspended in PBS + BSA, incubated at 4°C with saturating concentrations of the specific primary mouse monoclonal antibodies H52 (30 /~g/ml), Bear-1 (2.0 /zg/ml) and anti-Leu-8 (0.5 /zg/ml) which were used to recognize the adhesion molecules CD18, C D l l b and L-selectin, respectively. An irrelevant mouse IgG 1 (30 /zg/ml) primary antibody was used as a negative control. The PMN suspension was then washed in PBS + BSA, and similarly incubated, in the dark, with a fluorescent secondary antibody (5 tzg/ml of goat-antimouse IgG R-phycoerythrin-conjugated). After washing with PBS + BSA, labeled cells were either immediately analyzed on the flow cytometer or fixed with 1% paraformaldehyde in PBS for later analysis. An EPICS flow cytometer (Coulter Corporation; Hialeah, FL) was used with excitation wavelength of 488 nm. Gating based on the forward and side scatter was performed to isolate a homogeneous PMN population, excluding aggregates and debris. A frequency distribution histogram of the fluorescent signal was obtained from at least 5000 cells for each sample. Mode fluorescence was used as an index of light intensity and within each treatment group, the fluorescence of the IgG~-labeled sample was subtracted to correct for the non-specific antibody binding. 2.4. PMN-endothelial adhesion assays

Cultures of human aortic endothelial cells (HAECs; Clonetics Corporation; San Diego, CA) were grown in 6-well plates to near confluence at 37°C in a 5% CO 2 incubator [25]. Endothelial cell growth medium (EGM) was supplemented with the following growth factors and ingredients: bovine brain extract (12 /zg/ml), human recombinant epidermal growth factor (10 ng/ml), 2% fetal bovine serum, hydrocortisone (1.0 /zg/ml), gentamicin (50 /zg/ml) and amphote.ricin-B (50 ng/ml). Before the adhesion measurement, tumor necrosis factor (TNF)-a (10 ng/ml) wa,; added to predefined wells for a 4-h incubation period, followed by two washings with cytokine-free growth medium. Aliquots of isolated PMNs, suspended in PAG-CM, were exposed to 5.0 mM H20 z for 30 rain, as described above, and washed twice in EGM. Since H202 can up-regulate endothelial cell adhesion molecules [26], care was taken to avoid carry-over effect of this oxidant by repeated washings of treated PMNs prior to addition to endothelial cells. Untreated and H 202-treated

253

PMNs, in aliquots of 8 × 105 cells, were pipetted separately into respective wells containing endothelial monolayers. After 30 min incubation (37°C, 5% CO2), the supernatants were aspirated and the wells rinsed twice with PBS to remove non-adherent PMNs. PMNs that remained adherent to endothelial cells were counted under a phasecontrast inverted light microscope (Nikon Diaphot ®) at a 100 × magnification on a Universal Imaging system with a Sony Trinitron ® screen. Three wells were studied for each condition and three microscopic fields per well were counted. The data were normalized to basal values of the adherence of untreated PMNs to untreated HAECs. To compare the effect of H202 to a known chemotactic factor, similar adhesion assays were performed with PMNs pre-treated with human recombinant complement fragment 5a (C5a; 500 ng/ml). 2.5. Electron paramagnetic resonance measurements

Electron paramagnetic resonance (EPR) spectra were recorded in a fiat cell at room temperature with a BrukerIBM ER 300 spectrometer operating at X-band using a TMI~ 0 cavity, a modulation frequency of 100 kHz, modulation amplitude of 0.5 G, microwave power of 20 mW, and microwave frequency of 9.77 GHz. The digital spectral data were transferred to a personal computer for analysis. Software capable of isotropic spectral simulation was used for component analysis of experimental spectra as described previously [27]. 2.6. Chemicals and reagents

The chemicals and reagents used were purchased from the sources indicated: PIPES, contents of the PIPES buffer, xanthine, xanthine oxidase, BSA fraction V, H202, paraformaldehyde, iron, NTA, bovine erythrocyte superoxide dismutase 4000 units/mg, and C5a (Sigma Chemicals, St. Louis, MO); Percoll (Pharmacia LKB, Biotechnology AB, Upsalla, Sweden); catalase 65 000 units/mg (Boehringer Mannheim, Mannheim, Germany); deferoxamine methanesulfonate (Ciba-Geigy, Basle, Switzerland); Dulbecco's PBS (Gibco Laboratories, Grand Island, NY); mouse IgG 1 and Bear-1 (AMAC, Westbrook, ME), Anti-Leu-8 (Becton-Dickinson Immunocytometry Systems, San Jose, CA); goat-anti-mouse IgG R-phycoerythrin-conjugated (Tago, Burlingame, CA); H52 (gift from Dr. James Hildreth, Johns Hopkins University) DMPO (Aldrich Chemicals, Milwaukee, WI); and, TNF-a (Biosource International, Camarillo, CA). Double-distilled, deionized water was used to prepare all of the solutions. 2.7. Statistics

Data are presented as mean 5: S.E.M. Statistical analysis of the dose-response and time course data was performed by single and two-factor analysis of variance

A. Fraticelli et a l . / Biochimica et Biophysica Acta 1310 (1996) 251-259

254

(ANOVA), respectively. A P-value less than 0.05 was assumed to indicate statistical significance. When ANOVA revealed a statistical significance, and for all other data, paired or unpaired Student's t-test was used.

f 160] .~ 120t OO = o

3. Results

*~8' .

".

' 80

--o-- CD11 b

m

3.1. Alterations of PMN adhesion molecule expression after oxidant stimulation Flow cytometry experiments were performed on PMNs exposed to H 2 0 2. After H 2 0 2 exposure, up-regulation of the adhesion molecules CD18 and C D l l b was observed with a clear shift to the right seen in the flow cytometry histograms from these cells. Down-regulation of L-selectin was seen in these cells with a marked shift to the left in the histogram (Fig. 1). Increasing concentrations of H 2 0 2 resulted in a concentration-dependent increase in C D l l b

I - - - I Untreated PMNs I I H202 (1 mM) - treated PMNs

CD18

CD11b

= = O

L) ,

. . . .

.'.~';,"~"

. . . . . . .

Jkk ,

,..-.

......

40 u.

0.1

.

.

.

.

.

.

.

.

i

.

.

.

.

.

1 H;O2Concentration(raM)

.

.

.

i

10

Fig. 2. Effect of varying concentrations of H202 on PMN adhesion molecule expression. Values are expressed in % of mode fluorescence intensity of control (CT), untreated PMNs. Data are plotted as mean + S.E.M. of 6 experiments and are compared by two-way analysis of variance. * P _< 0.05: untreated versus treated PMNs.

and CD18 on the surface of PMNs, while the expression of L-selectin was decreased (Fig. 2). When compared to control, the lowest dose tested (0.1 mM) demonstrated small but statistically significant changes in CD18 (6.2 + 3.6%) and L-selectin expression (12.0 + 2.6%). For CDI lb expression, the threshold dose for a significant effect was 0.5 raM. At the highest dose tested (10 mM), H202 completely abolished L-selectin expression (0.19 _+ 0.07%), while CD18 and C D l l b expressions were increased by 73.1 _+ 11.8% and 60.8 _+ 18.1%, respectively. The time course of the effects of H202 on adhesion molecule expression was studied. The CD18 and C D l l b expression exhibited parallel changes with an early decrease after one minute exposure followed by a marked increase over the period between 1 and 15 min, after which only gradual further increases were seen toward plateau values after 30 min (Fig. 3). At 60 min, C D l l b and CD18

2OO ,

L-Selectin

~o 120 ~- ~

/;

~ 80 O

=

% ,~,L

• ~1; '~

%

.

........

"% ,..,...,..,_,.,.,,

Fluorescence Intensity (Log Scale) Fig, 1. Flow cytometry histograms from PMNs, stimulated for 30 min with 1.0 mM H202, or from untreated PMNs. In the top panel, up-regulation of CD18 is seen in H202-treated PMNs with a rightward shift of the fluorescence intensity histogram. On the center panel, a similar response is seen for C D l l b surface expression. In the lower panel, a down-regulation of L-selectin expression is noted.

I

40 0

•

CD18

•

L-Selectin

1"0 ;~0 30 i0 gO 6"0 Time (Minutes)

Fig. 3. Kinetics of H202-induced alterations in adhesion molecule expression of PMNs exposed to 5.0 mM H202. Within 1 rain of exposure to H202, L-selectin is down-regulated from the cell surface, this is followed by Mac-1 (CD11b/CD18) up-regulation after 15 rain of exposure. Values are expressed as % of mode fluorescence intensity of control (CT), untreated PMNs, plotted as mean + S.E.M. of 4-5 experiments, and compared by two-way analysis of variance. * P _< 0.05: untreated versus treated PMNs.

255

A. Fraticelli et al. / Biochimica et Biophysica Acta 1310 (1996) 251-259

mode fluorescence intensities had increased from 102.7 + 25.7 to 200.8 +__32.7 ( P < 0.05) and from 129.3 + 26.7 to 247.1 _ 15.5 ( P < 5 X 10 - 5 ) respectively. In untreated PMNs no significant up-n~gulation of CD18 or C D l l b ; was observed. L-selectin e:rdaibited an opposite pattern: it was rapidly down-regulated following 1 min exposure of H 2 0 2, with a decrease in mode fluorescence from 66.1 _ 8.8 to 38.1 ___ 16.8 after 1 min and to 1.6 ___0.3 after 15 min of exposure, with no significant change afterwards ( P < 5 × 10-7)(Fig. 3). Control PMNs had an L-selectin fluorescence that remained constant throughout the observation period. In order to determine if the effects of HeO 2 on adhesion molecule expression were mediated by the formation of the reactive hydroxyl radical, EPR spectroscopy was applied to measure radical generation in these cells. In the presence of the spin trap DMPO at a concentration of 50 mM no significant signal was observed from PMNs alone or in the presence of 1.0 mM H 2 0 2 (Fig. 4A,B). On addition of the iron chelate Fe3+-NTA, however, a prominent 1:2:2:1 quartet signal, hyperfine coupling a H = a N = 14.9 G of DMPO-OH, was seen, secondary to the trapping of • OH as previously reported (Fig. 4C) [21]. Addition of 1 m M deferoxamine quenched this radical generation (Fig. 4D). Thus, addition of H:,O 2 to PMNs did not result in significant • OH generation in the absence of added iron. To further study whether inducing • OH formation would significantly amplify the effects of H 2 0 2 alone on adhesion molecule expression, experiments were performed incubating PMNs in the presence of the • OH generating

A

• 804

CD18

[ ] CD11b

.

•

L-Selectin

t.

,

(R C m i-

H~O2

o~

[

o

Fig. 5. Modulationof PMN adhesion molecule expression by H202 or hydroxyl radical (. OH) generation. H202 (1.0 mM) and •OH (1.0 mM of HzO2 and 20 /xM of ferric iron chelate Fe3+-nitrilotriacetate). % change in mode fluorescence intensity is plotted as mean+ S.E.M. of 4 experiments. * P < 0.05: untreatedversus treated PMNs.

system consisting of 1.0 mM H 2 0 2 and 2 0 / z M Fe3÷-NTA. No significant effect was seen in the presence of the Fe3+-NTA beyond that observed with H 2 0 2 alone (Fig. 5). CD1 l b and CD18 up-regulation was seen with L-selectin down-regulation but the magnitude of this process was identical to that with H 2 0 2 alone. This suggests that enhancing • OH generation in the PMNs does not potentiate the action of H 2 0 2, To further determine the role of H 2 0 2 verses • OH in the modulation of adhesion molecule expression, experiments were performed to determine if either deferoxamine or catalase quenched the effects of H 2 0 2 on CD18, C D l l b , and L-selectin expression. As expected, 500 U / m l of catalase almost totally quenched the effects of 5.0 mM H202; 1.0 mM deferoxamine, however, had no significant effect (Fig. 6). These experi-

B • CD18

•

L-Selectin

~" 40

4

~o-~ o

D

-40.

HK)2 + Catalase

o ,

3435

[ ] CD11 b

I 3460

,

I 3485

,

I 3510

n-, -80.

,

3535

MagneticField(Gauss) Fig. 4. EPR spectra of 1 × 10 6 PMNs/ml in the presence of 50 mM DMPO. (A) PMNs in the presence of DMPO alone; (B) with additionof 5 mM H202; (C) with addition of 5 mM H202 and 20 /zM Fe3+-NTA; (D) as in (C) but with 1.0 mM deferoxaminealso added. Each spectrum is a sum of 4, 1-min acquisitions. In the absence of H202 no radical generation was observed, while in the presence of HeO2 alone again no significant radical generation was seen. In the presence of Fe3+-NTA, however, marked radical generation occurred with the formation of a large DMPO-OH signal, and this radical generation was quenched by deferoxamine.

H202+ mira Deferoxamine

Fig. 6. Modulationof PMN adhesion moleculeexpressionby exposure to H202 in the presence of catalase (500 units/ml) or deferoxamine (1.0 mM). Catalase, a specific H202 scavenger,markedlyblocked the effects of H202 on adhesionmoleculeexpression. The effect of H202 (5.0 mM) on adhesion molecule expression does not appear to be due to the iron-catalyzedproductionof hydroxyl radicals, since the changes induced by H20 z in the presence of the iron chelator deferoxaminewere similar to that inducedby H~O2 alone. Adhesionmoleculeexpressionis plotted as in Fig. 5 for 4 experiments. " P < 0.05: untreated versus treated PMNs.

256

A. Fraticelli et al. / Biochimica et Biophysica Acta 1310 (1996) 251 259

A "~ I"

5°t 4O 30

L-Selectin

~u -20 "10 t -30

,"T

-40

CDllb

[ ] Xanthine 0.4 mM [ ] Xanthine 0.8 mM [ ] Xanthine 1.2 mM

* -T *

-50

C L 3435

CD18

3460

,

I 3485

,

I 3510

, 3535

Magnetic Field (Gauss) Fig. 7. EPR spectra of 1 × 10 6 PMNs/ml in the presence of 50 mM DMPO. (A) PMNs in the presence of DMPO alone. (B) with addition of xanthine oxidase 40 mU/ml and xanthine 0.4 mM. (C) as in (B) but with the addition of 200 U / m l of superoxide dismutase. Each spectrum consists of the sum of 4, l-min acquisitions. In the presence of DMPO alone no radical generation is seen while in the presence of xanthine oxidase and xanthine a prominent signal is seen consisting largely of the DMPO-superoxide adduct, DMPO-OOH, with a small component of DMPO-OH.

m e n t s f u r t h e r s u g g e s t t h a t the effects o f H 2 0 2 are n o t hydroxyl radical-mediated. In o r d e r to d e t e r m i n e if the s u p e r o x i d e free radical also i n d u c e d c h a n g e s in P M N a d h e s i o n m o l e c u l e e x p r e s s i o n ,

Fig. 8. Effect of superoxide on PMN adhesion molecule expression. PMNs were exposed to a varying magnitude of superoxide generation by varying the concentration of the substrate xanthine. CD18, CDI lb, and L-selectin surface expression responded to superoxide stimulation in a dose-dependent manner. Values are expressed in percent changes of mode fluorescence intensity of control, untreated PMNs. Data are plotted as mean_S.E.M, of 4 experiments and are compared by Student's t-test. * P < 0.05, * * P < 0.01: untreated versus treated PMNs.

e x p e r i m e n t s w e r e p e r f o r m e d e x p o s i n g P M N s to a specific superoxide generating system consisting of xanthine and x a n t h i n e o x i d a s e in the p r e s e n c e o f c a t a l a s e a n d d e f e r o x a m i n e . E P R e x p e r i m e n t s in the p r e s e n c e o f the spin trap D M P O , 5 0 m M , s h o w e d t h a t P M N s a l o n e did n o t give rise to s i g n i f i c a n t radical g e n e r a t i o n , w h i l e in the p r e s e n c e o f x a n t h i n e a n d x a n t h i n e o x i d a s e a p r o m i n e n t signal was o b s e r v e d p r i m a r i l y d u e to the D M P O - O O H a d d u c t o f t r a p p e d s u p e r o x i d e w i t h h y p e r f i n e c o u p l i n g s a s = 14.2 G

Fig. 9. Photomicrographs of PMN-endothelial adhesion assays. Unstimulated PMNs and unstimulated monolayers of human aortic endothelial cells (CONTROL), hydrogen peroxide (5.0 mM)-stimulated PMNs and unstimulated endothelial cells (H202), complement 5a (500 ng/ml)-stimulated PMNs and unstimulated endothelial cells (C5a), and tumor necrosis factor-or (10 ng/ml)-stimulated endothelial cells and unstimulated PMNs (TNF). When the PMNs were stimulated with H202, a marked increase in the number of adherent PMNs occurred compared to unstimulated PMNs. This increase in PMN adhesion is similar to that observed with the PMN activator C5a or when the endothelial cells were stimulated with TNF.

258

A. Fraticelli et al. / Biochimica et Biophysica Acta 1310 (1996) 251-259

ited similar effects with clear dose-dependent up-regulation of CD18 and CD1 lb observed. The concentrations of xanthine and xanthine oxidase utilized to generate superoxide in these experiments were of the same order of magnitude as those which have been shown to occur in the ischemic rat heart [37]. Therefore, it is possible that the H202 and superoxide-dependent effects observed in these in vitro studies could also occur within postischemic cells and tissues. In addition to being triggered by endothelialderived radical generation, these changes could be further triggered by the high levels of oxidants which can be generated by the NADPH oxidase of the PMN itself after PMN activation by other chemotactic factors. This phenomenon may represent a feedback amplification loop whereby activated PMNs further trigger recruitment, adhesion, and activation of other PMNs, which would in turn result in further oxidant formation. A novel observation in our study is that H202 or superoxide caused a loss of surface expression of the adhesion molecule L-selectin. Again significant changes were present at a H 2 0 2 concentration of 0.1 mM, with maximum effect seen above 1 mM. At mM concentrations time course experiments showed that a 33% decrease occurred after 1 min of exposure and that 5 rain were required to complete this process. This is similar to the rapid down-regulation of PMN L-selectin observed after exposure to C5a [34]. L-Selectin mediates early adhesion of PMNs to stimulated endothelium under flow conditions [34], but not PMN extravasation, an integrin-dependent process [35]. L-Selectin shedding is believed to play an important role in PMN activation, since it may serve as a mechanism triggering transendothelial migration [34,35]. Therefore, our finding of L-selectin down-regulation, accompanied by integrin up-regulation, supports the hypothesis of a direct chemotactic effect of H 2 0 2 on human PMNs. A series of experiments were performed to determine whether the effects of H202 were mediated by the formation of the hydroxyl free radical. In other cell types such as myocytes, it has been observed that the cytotoxicity of H202 is due to the iron-mediated reaction to form • OH. In human PMNs, however, little if any • OH generation was observed on addition of H202 as demonstrated by EPR spin trapping. Furthermore the high affinity ferric iron chelator deferoxamine which is known to block the ironmediated Fenton reaction did not block the effects of H202 on PMN adhesion molecule expression. On addition of the iron chelate Fe3+-NTA to PMNs in the presence of H202, marked .OH generation was observed as demonstrated by the formation of the DMPO-OH adduct in EPR spin trapping experiments. No enhancement of the effects of H202 on adhesion molecule expression, however, was observed in the presence of Fe3+-NTA, further suggesting that •OH formation did not further activate PMNs beyond the effect seen with H 2 0 2 alone. We observed that the H202-induced changes in PMN

adhesion molecule expression are associated with significantly increased adhesion to human endothelial cells. The magnitude of this cellular adhesion was comparable to that observed upon pretreatment with the known PMN activator C5a. The increased adhesion observed is probably attributable to Mac-l up-regulation, as our flow cytometry data showed that surface L-selectin was lost shortly after H202 exposure. This increase in adhesion was blocked by a specific CD18 monoclonal antibody, further suggesting that the adhesion was CD18-mediated. The pro-adhesive effect of H202 on PMNs was further enhanced when HAEC were prestimulated with TNF-a, which promotes the endothelial expression of ICAM-1, the counter receptor for Mac-1. Thus, H202 or the other oxidants and oxygen free radicals which we have observed to up-regulate the surface expression of CD1 lb and CD18 could amplify the magnitude of leukocyte-endothelial adhesion which occurs in response to other known stimulators of adhesion such as the cytokines. In summary, we observe that oxidants and oxygen-derived free radicals can directly modulate the expression of the adhesion molecules C D l l b , CD18, and L-selectin on the surface of human PMNs. These changes are accompanied by increased adhesion to resting and cytokine-stimulated human endothelial cells. This phenomenon could have a pathophysiologic role in the injury which occurs in reperfused tissues, where an endothelial-derived oxidant burst is followed by leukocyte chemotaxis and activation [38]. It could also be of great importance in a variety of other inflammatory disease processes in which initial leukocyte activation occurs resulting in oxidant formation which could in turn trigger an amplification loop of further leukocyte chemotaxis, adhesion, and activation. Further studies in physiological models will be required to confirm the relevance and importance of these observations. Since our studies were performed on human cells, these data suggest that oxidants and oxygen free radicals may modulate the regulation of leukocyte-endothelial interactions in humans and may play an important role in the pathophysiology of inflammation in human disease.

Acknowledgements The authors thank Dr. James E. Hildreth (Department of Pharmacology and Experimental Therapeutics, The Johns Hopkins Medical Institutions) for generously providing the monoclonal antibody H52. We also wish to thank Dr. Stefano Corda for supplying cultured human aortic endothelial cells used in the adhesion assays. This work was supported by the National Institutes of Health Grants HL-52315 and HL-38324. J.L.Z. is also supported by an Established Investigator Award from the American Heart Association.

A. Fraticelli et al. / Biochimica et Biophysica Acta 1310 (1996) 251-259

References [1] Jolly, S.R., Kane, W.J., Bailie, M.B., Abrams, G.D. and Lucchesi, B.R. (1984) Circ. Res. 54, 277-285. [2] Zweier, J.L., Kuppusamy, P., Williams, R., Raybum, B.K., Smith, D., Weisfeldt, M.L. and Flaberty, J.T. (1989) J. Biol. Chem. 264, 18890-18895. [3] Kraemer, R.B., Seligmann, I3. and Mullane, K.M. (1990) Am. J. Physiol. 258, H1847-1855. [4] Lucchesi, B.R. (1990) Annu. Rev. Physiol. 52, 561-576. [5] Jackson, R.L. (1993) Biochern. Soc. Trans. 21,650-651. [6] Crawford, D.W. (1991) Atherosclerosis 89, 97-108. [7] Bevilacqua, M.P. (1993) Annu. Rev. Immunol. 11,767-804. [8] Gallatin, W.M., Weissman, I.L. and Butcher, E.C. (1983) Nature 304, 30-34. [9] yon Andrian, U.H., Chamber:~, J.D., McEvoy, L.M., Bargatze, R.F., Arfors, K.E. and Butcher, E.C. (1991) Proc. Natl. Acad. Sci. USA 88, 7538-7542. [10] Watson, S.R., Fennie, C. znd Lasky, L.A. (1991) Nature 349, 164-167. [11] Harlan, J.M. (1985) Blood 65, 513-525. [12] Rothlein, R., Dustin, M.L., Marlin, S.D. and Springer, T.A. (1986) J. Immunol. 137, 1270-1274. [13] Simmons, D., Makgoba, M.W. and Seed, B. (1988) Nature 331, 624-627. [14] Zweier, J.L., Kuppusamy, P. and Lutty, G.A. (1988) Proc. Natl. Acad. Sci. USA 85, 4046-4(150. [15] Zweier, J.L., Broderick, R., Kuppusamy, P., Thompson-Gorman, S. and Lutty, G.A. (1994) J. Biol. Chem. 269, 24156-24162. [16] Ratych, R.E., Chuknyiska, R.S. and Bulkley, G.B. (1987) Surgery 102, 122-131. [17] Grisham, M.B., Hemandez, L.A. and Granger, D.N. (1986) Am. J. Physiol. 251,567-574. [18] Granger, N.D. Benoit, J.N., Suzuki, M. and Grisham, M.B. (1989) Am. J. Physiol. 257, G683-689. [19] Suzuki, M., Inauen, W., Kvietys, P.R., Grisham, M.B., Meininger, C., Schelling, M., Granger, H.J. and Granger, D.N. (1989) Am. J. Physiol. 257, H1740-1745. [20] Schleimer, R.P. and Rutledge, B.K. (1986) J. Immunol. 136, 649654.

259

[21] Josephson, R.A., Silverman, H.S., Lakatta, E.G., Stem, M.D. and Zweier, J.L. (1991) J. Biol. Chem. 266, 2354-2361. [22] Kuppusamy, P. and Zweier, J.L. (1989) J. Biol. Chem. 264, 98809884. [23] Weiss, S.J., Klein, R., Slivka, A. and Wie, M. (1982) J. Clin. Invest. 70, 598-607. [24] Seed, J.L., Kensler, T.W., Elia, M. and Trush, M.A. (1990) Res. Comm. Chem. Pathol. & Pharm. 67, 349-360. [25] Bochner, B.S., McKelvey, A.A., Schleimer, R.P., Hildreth, J.E.K. and MacGlashan Jr., D.W. (1989) J. Immunol. Methods 125, 265271. [26] Zweier, J.L. (1988) J. Biol. Chem. 263, 1353-1357. [27] Patel, K.D., Zimmerman, G.A., Prescott, S.M., McEver, R.P. and Mclntyre, T.M. (1991) J. Cell. Biol. 112, 749-759. [28] Palluy, O., Morliere, L., Gris, J.C., Bonne, C. and Modat, G. (1992) Free Rad. Biol. Med. 13, 21-30. [29] Lewis, M.S., Whatley, R.E., Cain, P., Mclntyre, T.M., Prescott, S.M. and Zimmerman, G.A. (1988) J. Clin. Invest. 82, 2045-2055. [30] Patel, K.D., Zimmerman, G.A., Prescott, S.M. and Mclntyre, T.M. (1991) J. Biol. Chem. 266, 11104-11110. [31] Suzuki, M., Asako, H., Kubes, P., Jennings, S., Grisham, M.B. and Granger, D.N. (1991) Microvasc. Res. 42, 125-138. [32] Petrone, W.F., English, D.K., Wong, K. and McCord, J.M. (1980) Proc. Natl. Acad. Sci. USA 77, 1159-1163. [33] Skoglund, G., Cotgreave, I., Rincon, J., Patarroyo, M. and Ingelman-Sundberg, M. (1988) Biochem. Biophys. Res. Commun. 157, 443 -449. [34] Krishimoto, T.K., Jutila, M.A., Berg, E.L. and Butcher, E.C. (1989) Science 245, 1238-1241. [35] Smith, C.W., Kishimoto, T.K., Abbassi, O., Hughes, B.J., Rothlein, R., Mclntire, L.V., Butcher, E. and Anderson, D.C. (1991) J. Clin. Invest. 87, 609-618. [36] Klebanoff, S.J. (1988) Inflammation: Basic Principles and Clinical Corerelates (Gallin, J.l., Goldstein, I.M. and Snyderman, R., eds.), Raven Press, New York, pp. 391-444. [37] Thompson-Gorman, S.L. and Zweier, J.L. (1990) J. Biol. Chem. 265, 6656-6663. [38] Shandelya, S., Kuppusamy, P. and Zweier, J.L. (1993) Circulation 87, 536-546.