Mechanisms of human neutrophil oxidant production after severe injury

Mechanisms of human neutrophil oxidant production after severe injury Gina Quaid, MD, Cindy Cave, BA, Mark A. Williams, PhD, Robert F. Hennigan, PhD, Gary Bokoch, PhD, and Joseph S. Solomkin, MD, Cincinnati, Ohio, and La Jolla, Calif

Background. The purpose of this study was to determine the mechanisms of enhanced oxidant production after severe injury. Methods. Neutrophils were harvested from patients within 24 hours of admission who had an injury severity score greater than 16. Nonadherent and adherent neutrophil oxidant production was measured after N-formyl-methionyl-leucyl-phenylalanine (fMLP) stimulation. Translocation of cytochrome b558 and cytosolic components p47phox and p67phox were determined by oxidation-reduction spectroscopy and immunoblotting, respectively. Flow cytometry measured integrin expression. Integrin and p47phox colocalization was examined by confocal microscopy. Results. Eighteen patients were studied within 15 ± 1.4 hours. Four women and 14 men suffered a blunt injury and had a mean injury severity score of 22 (range, 16 to 34). Nonadherent patient neutrophils showed a decrease in fMLP-stimulated oxidant production, whereas adherent neutrophil oxidant production was increased in both the vehicle control and fMLP-stimulated groups. Nicotinamide adenine dinucleotide phosphate (NADPH) oxidase components p47phox and cytochrome b558 were mobilized to the plasma membrane, whereas p67phox showed minimal change. Integrin CD11b α chain showed a significant increase in expression. Confocal microscopy showed colocalization of p47phox and α chain CD11b on the plasma membrane of patient neutrophils. Conclusions. Colocalization of NADPH oxidase components and integrins may regulate the enhanced oxidant production in human neutrophils after severe injury. (Surgery 2001;130:669-76.) From the Department of Surgery, University of Cincinnati, College of Medicine, the Department of Cell Biology, University of Cincinnati, College of Medicine, The Vontz Center for Molecular Studies, Cincinnati, Ohio, and the Department of Immunology, Research Institute of Scripps Clinic, La Jolla, Calif

DESTRUCTION OF END ORGANS after severe injury is thought to be secondary to heightened neutrophil functions. Primed neutrophils isolated from trauma patients display increased oxidative response as well as an increased ability to sequester in end organs.1,2 The biochemistry by which these functions are regulated is yet to be fully elucidated. Oxidant production occurs through formation of a membrane-associated complex, the nicotinamide adenine dinucleotide phosphate (NADPH) oxidase. The enzyme components are separated into membrane (cytochrome b558 and

Presented at the 58th Annual Meeting of the Central Surgical Association, Tucson, Ariz, March 7-10, 2001. Supported by NIH grant #T12GM08478-04, Shriner’s Hospitals for Children grant #8690, and NIH RO1 GM44428. Reprint requests: Gina A. Quaid, MD, Department of Surgery, University of Cincinnati, ML #558, Cincinnati, OH 45267-0558. Copyright © 2001 by Mosby, Inc. 0039-6060/2001/$35.00 + 0 11/6/116923 doi:10.1067/msy.2001.116923

Rap1A) and cytosolic (p47phox, p67phox, p40phox, and p21rac) components. On activation, the cytosolic components translocate to the plasma membrane and assemble with membrane-bound components, resulting in the active NADPH oxidase. The molecular basis for priming of suspension phase cells is not known. In vitro studies have revealed a stimulus specificity: partial mobilization of components to the plasma membrane (with lipopolysaccharide stimulation) or partial phosphorylation of the p47phox component without translocation (with granulocyte-macrophage colony-stimulating factor stimulation).3,4 Sequestration is regulated by activation of cell surface integrins. Integrins are ubiquitous transmembrane adhesion molecules. They are also able to regulate intracellular signaling pathways.5 Perhaps the most destructive response induced by ligation of activated integrins is the generation of potent oxidants. The current studies were undertaken to examine the response of circulating patient neutrophils in assays specifically examining adherSURGERY 669

670 Quaid et al

ence-dependent (integrin-signaled) oxidant production. We also examine the molecular basis for primed oxidant production in the nonadherent neutrophils. METHODS Patient selection process. Trauma patients were studied within 24 hours of admission to the hospital. This was chosen to allow for patient stabilization before neutrophil isolation. All patients with head injuries were excluded. Patients who received multiple transfusions and/or had an injury severity score (ISS) greater than 16 were included in this study. This study was reviewed and approved by the Institutional Review Board at the University of Cincinnati. Neutrophil isolation. After obtaining informed consent from patients and healthy volunteers, venous blood was drawn in syringes containing 15% ethylenediamine tetraacetic acid (Sigma, St Louis, Mo). Red blood cells were removed by dextran (Pharmacia, Piscataway, NJ) sedimentation. The leukocyte rich supernatant was layered over Histopaque 1077 (Sigma) and centrifuged at 400g for 30 minutes at 4°C. Pelleted neutrophils were resuspended in phosphate-buffered saline (PBS) (Gibco, Grand Island, NY), and residual red blood cells were removed by hypotonic lysis. Neutrophil preparations were routinely greater than 95% pure with greater than 95% viability by trypan blue exclusion. Kinetic superoxide assay. The rate of superoxide production by neutrophils was measured by the superoxide dismutase–inhibitable reduction of ferricytochrome C. Neutrophils (2 × 105) were added to a 96-well plate containing 75 µmol/L ferricytochrome C (Sigma), ± 60 µg/mL superoxide dismutase (Sigma), and warmed to 37°C. Buffer, phorbol myristate acetate (PMA, 100 ng/mL), N-formyl-methionyl-leucyl-phenylalanine (fMLP, 1 µmol/L) were added, and the plate was read at 550 nm in kinetic mode at 2-minute intervals for 10 minutes with a ThermoMax microplate reader (Molecular Devices, Menlo Park, Calif). The rate of superoxide production was calculated from changes in milli optical density/minute by using an extinction coefficient of 20 × 10–3 (OD) µmol/L–1 cm–1. Adherence-dependent oxidative response assay. Primaria 96-well polystyrene tissue culture dishes were coated with 1 µg/well fibronectin for 2 hours at 37°C and 5% CO2 and then washed. Each well contained 100 µL of reaction mixture (10 mmol/L scopoletin, 1 mg/mL horseradish peroxidase, 4 mmol/L NaN3 in Krebs-Ringer phosphate buffer plus glucose [KRPG]) and 20 µL neutrophils resus-

Surgery October 2001 pended in KRPG at 7.5 × 105 cells/mL that were allowed to incubate 15 minutes at 37°C before stimulation with buffer, fMLP (10 nmol/L), or PMA (50 ng/mL). Fluorescence was measured immediately and at 10-minute intervals for 80 minutes. H2O2 production was calculated from the decrease in fluorescence due to the oxidation of scopoletin. The data are expressed as nanomoles of H2O2 produced per 1.5 × 104 polymorphonuclear neutrophils. Cell fractionation by N2 cavitation. Normal and patient neutrophils were treated with 5 mmol/L diisopropyl fluorophosphate for 15 minutes followed by isolation of subcellular fractions as previously described.3 The membrane and granule bands were aliquoted and stored at –80°C until use. Spectral analysis of flavocytochrome b558. Membrane fractions (30 µg protein) isolated from trauma patients and normal donors as described previously were lysed with 1% Triton-X 100 and scanned (Du 640 Spectrophotometer; Beckman Coulter, Fullerton, Calif) between 600 and 400 nm to obtain baseline (oxidized state). The amount of flavocytochrome b558 was calculated by subtracting the reduced peak from the oxidized peak at 558 nm and using an extinction coefficient of 21.6 mmol/L–1 cm–1. Immunoblotting technique. Two to five micrograms of membrane protein was resolved on 10% sodium dodecylsulfate–polyacrylamide gel electrophoresis gel (Bio Rad, Hercules, Calif), transferred to a nitrocellulose membrane, and blotted with polyclonal antibodies directed against the oxidase components p47phox or p67phox (a gift from Gary Bokoch, Scripps Institute). The membranes were then exposed to goat anti-rabbit antibody (Pierce) conjugated to horseradish peroxidase, developed by using chemiluminescence reagents (New England Nuclear, Boston, Mass), and exposed to autoradiographic film. Radiographs were imaged and analyzed by Alpha Imager 2200 V5.04 (Alpha Imotech Corp, San Leandro, Calif) and expressed in arbitrary optical density units called integrated density volume (IDV). Flow cytometry. Two million freshly isolated neutrophils were resuspended in PBS containing 2% goat serum (Gibco) and incubated with monoclonal mouse anti-human integrin α chain CD11b, β chain CD18, or isotype control antibodies (Pharmingen, San Diego, Calif). After washing, neutrophils were incubated with fluorescently labeled goat anti-mouse immunoglobulin (Kirkegaard and Perry Labs, Gaithersburg, Md). After 2 final washes, cells were resuspended in 1% paraformaldehyde and analyzed on a Coulter Epics XL Flow Cytometer (Coulter, Miami, Fla). Flow

Surgery Volume 130, Number 4

cytometry readings were performed after standardization with Standard-Brite beads (Coulter). Confocal microscopy. Freshly isolated neutrophils (2 × 106) were fixed with 500 µL of 4% paraformaldehyde for 30 minutes, washed with PBS, and permeabilized with 500 µL FACS permeabilizing solution (Becton/Dickinson, San Jose, Calif) for 6 minutes at room temperature. Cells were treated as above except the primary antibodies were anti-p47phox and anti-CD11b, and the secondary antibodies included either Alexa 546 (Molecular Probes, Eugene, Ore) labeled anti-rabbit or Alexa 488 labeled anti-mouse antibody. One hundred thousand labeled neutrophils were resuspended in PBS, cytospun (Cytospin 2; Shandon Lipshaw, Inc, Pittsburgh, Pa) at 3000 rpm for 5 minutes, and wet mounted with Gel/Mount (Blφmed Corp, Foster City, Calif). Slides were examined and analyzed on an LSM 510 laser scanning confocal microscopy (Carl Zeiss Inc, Microscopy Division, Thornwood, NY). Statistical analysis. Statistical testing was done with the computer program Instat (Graphpad Software, Inc). Laura James, MA, statistician at the Shriners’ Burns Institute, Cincinnati, was available to us for consultation. Results of patient studies were compared to results obtained with simultaneously studied neutrophils from normal volunteers. These data were then handled as unpaired samples. We analyzed patient data in minimum groups of 7 individual samples. Significance was set at P = .05. RESULTS Patient groups. The study population consisted of 4 women and 14 men who had mostly blunt trauma with a mean ISS of 22 (range, 16 to 34). They had an average age of 46 years (range, 17 to 79 years), white blood cell count of 13.35 (± 1.69), and required on average 3.5 (range, 0 to 15) units of blood products. Nonadherent superoxide production. Severely injured patients showed a notable decrease (P < .05) in the ability to produce superoxide in response to fMLP but not PMA when compared with normal control subjects (Fig 1). Adherence-dependent peroxide production. Patients who had been severely traumatized showed an enhanced response to fMLP stimulation, but they also showed a greatly enhanced ability to generate peroxide in response to vehicle control itself (Fig 2). Cytochrome b558, p47, and p67 expression. Patients exhibited a significant increase in the amount of plasma membrane-associated cytochrome b558 and p47phox compared with normal control sub-

Quaid et al 671

Fig 1. Nonadherent superoxide (SO) generation. Severely injured patients (PT) showed a significantly decreased rate of superoxide production when stimulated with fMLP compared with normal subjects (NML) (*P < .05). The rate of superoxide production was measured by superoxide dismutase-inhibitable reduction of ferricytochrome C after stimulation with either buffer or fMLP.

Fig 2. Adherence-dependent peroxide production. The patient neutrophils studied in Fig 1 were simultaneously examined in an adherence-dependent assay. Adherent neutrophils generated significant amount of peroxide in vehicle control group after severe injury (PT) (P < .05) compared with normal control subjects (NML). In response to fMLP, there was minimal enhancement in patient neutrophils. Adherence-dependent peroxide production was studied by allowing neutrophils to adhere on a fibronectin surface and then stimulating with buffer or fMLP (10 nmol/L). Generation of peroxide was measured as oxidation in scopoletin fluorescence and expressed as nanomoles H202/1.5 × 104 cells/60 minutes.

jects (Figs 3 and 4, respectively; P < .05). Although elevated, membrane-associated p67phox was not significantly different than in normal control subjects (Fig 5). Integrin expression. Previous studies suggested that β2 integrin α chain CD11b and β chain CD18 are significantly increased after traumatic insult.6-8

672 Quaid et al

Surgery October 2001

brane surface after severe injury. Fig 7 reveals that in severely injured patients p47phox (red) and CD11b (green) were in close proximity compared with normal control subjects (red arrows at yellow plaques). Controls revealed that there was no crossreactivity between the secondary antibodies for p47phox or CD11b (data not shown).9

Fig 3. Cytochrome b558 analysis by oxidation-reduction spectroscopy. Membrane-associated cytochrome b558 was significantly increased in severely injured patients compared with normal subjects (P = .0086). Patient plasma membranes were harvested by nitrogen cavitation followed by discontinuous Percoll gradient centrifugation. After bicinchoninic acid determination of protein content, thirty micrograms of membrane protein was analyzed by oxidation-reduction spectroscopy.

A. Immunoblots

B. Spot Densitometry

Fig 4. Western blot (A) and spot densitometry analysis of p47phox (B). Plasma membrane concentration of p47phox was significantly increased in patients (P) (P = .043) compared with normal (N) control subjects. IDV, Integrated density volume.

In this study the α chain was significantly increased in patients compared with normal control subjects (Fig 6; P = .039), but the β chain expression was not altered after severe injury. Colocalization. Confocal microscopy was used to determine whether NADPH oxidase components and β2 integrin colocalized on the plasma mem-

DISCUSSION Oxidant production by adherent neutrophils is a critical element in the pathogenesis of organ injury after acute disease. Therapeutic strategies that disrupt integrin-mediated attachment, antioxidant administration, or interference with activation and degranulation of the cells prevent acute organ injury in a variety of animal models.10-13 We therefore chose to characterize oxidative responses of trauma patient neutrophils in an adherence-dependent model. This model has been extensively explored in normal neutrophils. Integrin regulation of the oxidative response has been demonstrated by 3 key observations: (1) neutrophils obtained from patients with leukocyte adhesion deficiency do not respond to tumor necrosis factor (TNF) or fMLP when adherent; (2) soluble anti-β2 integrin antibodies inhibit the adhesiondependent stimulation of normal neutrophils; and (3) surface-bound anti-β2 integrin antibodies mimic the stimulatory effect of TNF triggered spreading and production of toxic oxygen derivatives.14 We found that patient cells in suspension treated with buffer or with fMLP produced an oxidative response similar to or below that seen with normal volunteer neutrophils. However, buffer-treated cells placed on fibronectin matrices had a markedly enhanced rate of oxidant production compared with controls. When the production of hydrogen peroxide in buffer-treated cells was subtracted from the fMLP-stimulated response, fMLP induced a normal response (data not shown). This would imply that the enhanced oxidant production is not due to fMLP stimulation alone, but that intracellular alterations have occurred. To investigate this further, we wanted to determine whether these changes were due to modification in the NADPH oxidase. We then performed immunoblot analysis of neutrophil membrane fractions for oxidase components p47phox and p67phox and oxidation-reduction spectroscopy for cytochrome b558. Our results indicated a substantial and significant up-regulation of both the cytochrome component and p47phox. We did not find up-regulation of p67phox. The priming pattern is similar to that induced by lipopolysaccharide treatment of normal neutrophils.3

Quaid et al 673

Surgery Volume 130, Number 4

A. Immunoblots

B. Spot Densitometry

Fig 5. Western blot (A) and spot densitometry analysis of p67phox (B). Plasma membrane concentration p67phox was not considerably altered in patients (P) compared to normal control subjects (N).

Fig 6. Integrin expression. Integrin CD11b was significantly increased in patients (PT) compared with normal control subjects (NML) (P = .039), but CD18 was not significantly altered after severe injury. Expression of β2 integrin components was measured by flow cytometry and expressed as mean channel fluorescence.

Integrin expression in this study was similar to previously reported studies.6-8 Our confocal data suggest that clustering of integrins and oxidase components occurs in these patient cells. Yan et al14-16 studied human neutrophils on fibrinogen matrices exposed to TNF-α. They identified colocalization of integrins, NADPH oxidase components, tyrosine phosphorylated proteins, and the protein tyrosine kinase p58fgr in human neutrophils, molecules that are potentially involved in the formation of a submembranous signaling complex. Our findings support the notion that

the priming process may occur at least in part by colocalizing integrins and NADPH oxidase components. Whether signaling molecules (kinases) also associate with this complex needs further study. The signaling pathways leading from integrin ligation to oxidase formation are not well understood and are important because of the potential for selective pharmacologic intervention. It is apparent in in vitro studies that tyrosine kinases are involved in this process.15-19 Furthermore, it appears that p47phox may be targeted by several kinase cascades in intact neutrophils and is therefore a key regulatory site.20-22 The surface dependence produced by the requirement for integrin ligation and signaling to induce oxidant production provides an extraordinary level of control and localization of the inflammatory response. This geographic localization of oxidase activation protects oxidants from extracellular oxidant scavengers and limits the “bystander” injury that might otherwise occur.23 Proteases are also protected from inhibitors by the exclusive microenvironment of the focal adhesion. Alterations in neutrophil oxidative function after trauma have been related to the intravascular release of various effectors of cell function, including C5a, platelet-activating factor, interleukin-8, and others.13,24,25 Our data would suggest that if this effect is due to intravascular mediator exposure, the effect is to cause partial oxidase assembly and integrin activation.

674 Quaid et al

Surgery October 2001

Fig 7. Confocal microscopy. (A), Normal confocal; (B), patient confocal. After severe injury, the p47phox component (red) and the integrin α chain CD11b (green) appeared to be in close proximity to each other (B). This was shown in a 1-µm optical section by yellow plaques on the plasma membrane of neutrophils (large red arrows), which do not exist in normal control subjects.

REFERENCES 1. Botha AJ, Moore FA, Moore EE, Kim FJ, Banerjee A, Peterson VM. Postinjury neutrophil priming and activation: an early vulnerable window. Surgery 1995;118:358-64. 2. Botha AJ, Moore FA, Moore EE, Sauaia A, Banerjee A, Peterson VM. Early neutrophil sequestration after injury: a pathogenic mechanism for multiple organ failure. J Trauma 1995;39:411-7.

3. DeLeo FR, Renee J, McCormick S, et al. Neutrophils exposed to bacterial lipopolysaccharide upregulate NADPH oxidase assembly. J Clin Invest 1998;101:455-63. 4. Dang PM, Dewas C, Gaudry M, et al. Priming of human neutrophil respiratory burst by granulocyte/macrophage colony-stimulating factor (GM-CSF) involves partial phosphorylation of p47(phox). J Biol Chem 1999;274:20704-8. 5. Williams MA, Solomkin JS. Integrin-mediated signaling in

Surgery Volume 130, Number 4

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

human neutrophil functioning. J Leukoc Biol 1999;65: 725-36. Giannoudis PV, Smith RM, Banks RE, Windsor AC, Dickson RA, Guillou PJ. Stimulation of inflammatory markers after blunt trauma. Br J Surg 1998;85:986-90. Maekawa K, Futami S, Nishida M, et al. Effects of trauma and sepsis on soluble L-selectin and cell surface expression of L-selectin and CD11b. J Trauma 1998;44:460-8. Botha AJ, Moore FA, Moore EE, Peterson VM, Goode AW. Base deficit after major trauma directly relates to neutrophil CD11b expression: a proposed mechanism of shockinduced organ injury. Intensive Care Med 1997;23:504-9. Mukherjee G, Rasmusson B, Linner JG, et al. Organization and mobility of CD11b/CD18 and targeting of superoxide on the surface of degranulated human neutrophils. Arch Biochem Biophys 1998;357:164-72. Mulligan MS, Smith CW, Anderson DC, et al. Role of leukocyte adhesion molecules in complement-induced lung injury. J Immunol 1993;150:2401-6. Lu H, Smith CW, Perrard J, et al. LFA-1 is sufficient in mediating neutrophil emigration in Mac-1- deficient mice. J Clin Invest 1997;99:1340-50. Mulligan MS, Wilson GP, Todd RF, et al. Role of beta 1, beta 2 integrins and ICAM-1 in lung injury after deposition of IgG and IgA immune complexes [published erratum appears in J Immunol 1993;150:5209]. J Immunol 1993;150:2407-17. Au BT, Williams TJ, Collins PD. Zymosan-induced IL-8 release from human neutrophils involves activation via the CD11b/CD18 receptor and endogenous platelet-activating factor as an autocrine modulator. J Immunol 1994;152:5411-9. Yan SR, Fumagalli L, Dusi S, Berton G. Tumor necrosis factor triggers redistribution to a Triton X-100- insoluble, cytoskeletal fraction of beta 2 integrins, NADPH oxidase components, tyrosine phosphorylated proteins, and the protein tyrosine kinase p58fgr in human neutrophils adherent to fibrinogen. J Leukoc Biol 1995;58:595-606. Yan SR, Fumagalli L, Berton G. Activation of SRC family kinases in human neutrophils: evidence that p58C-FGR and p53/56LYN redistributed to a Triton X-100- insoluble cytoskeletal fraction, also enriched in the caveolar protein caveolin, display an enhanced kinase activity. FEBS Lett 1996;380:198-203. Yan SR, Fumagalli L, Berton G. Activation of p58c-fgr and p53/56lyn in adherent human neutrophils: evidence for a role of divalent cations in regulating neutrophil adhesion and protein tyrosine kinase activities. J Inflamm 1995;45:297-311. Barry ST, Flinn HM, Humphries MJ, Critchley DR, Ridley AJ. Requirement for Rho in integrin signalling. Cell Adhes Commun 1997;4:387-98. Berton G, Fumagalli L, Laudanna C, Sorio C. Beta 2 integrindependent protein tyrosine phosphorylation and activation of the FGR protein tyrosine kinase in human neutrophils. J Cell Biol 1994;126:1111-21. Defilippi P, Venturino M, Gulino D, et al. Dissection of pathways implicated in integrin-mediated actin cytoskeleton assembly: involvement of protein kinase C, Rho GTPase, and tyrosine phosphorylation. J Biol Chem 1997;272: 21726-34. Waite KA, Wallin R, Qualliotine-Mann D, McPhail LC. Phosphatidic acid-mediated phosphorylation of the NADPH oxidase component p47-phox: evidence that phosphatidic acid may activate a novel protein kinase. J Biol Chem 1997;272:15569-78.

Quaid et al 675

21. Sue-A-Quan AK, Fialkow L, Vlahos CJ, et al. Inhibition of neutrophil oxidative burst and granule secretion by wortmannin: potential role of MAP kinase and renaturable kinases. J Cell Physiol 1997;172:94-108. 22. Zu YL, Qi J, Gilchrist A, et al. p38 mitogen-activated protein kinase activation is required for human neutrophil function triggered by TNF-alpha or FMLP stimulation. J Immunol 1998;160:1982-9. 23. Campbell EJ, Senior RM, Welgus HG. Extracellular matrix injury during lung inflammation. Chest 1987;92:161-7. 24. Kubes P, Niu XF, Smith CW, Kehrli ME Jr, Reinhardt PH, Woodman RC. A novel beta 1-dependent adhesion pathway on neutrophils: a mechanism invoked by dihydrocytochalasin B or endothelial transmigration. FASEB J 1995;9: 1103-11. 25. Rainger GE, Fisher AC, Nash GB. Endothelial-borne platelet-activating factor and interleukin-8 rapidly immobilize rolling neutrophils. Am J Physiol 1997;272:H114-22.

DISCUSSION Dr Mark A. Malangoni (Cleveland, Ohio). This is a very sophisticated study that analyzes the important components of the whole cycle of neutrophil dysfunction, and it has tremendous application in patients. I just have one comment and a question. I think it might help your study if you would include less severely injured patients as an additional control group. That might sort out whether it is the overall degree of injury or whether even a minor injury might contribute to some abnormalities in the neutrophil. The question that I had relates to the potential interventions. If these abnormalities are associated with organ injury, what type of interventions do you think will be important or what things do you suggest we try in the management of these patients? Dr James Tyburski (Detroit, Mich). A follow-up question to that is, have you looked at using any antibody, specifically anti-CD11 or anti-CD18, that has been tried clinically with very variable results? Dr Quaid. In regards to studying less severely injured patients, we previously studied all types of injury during our exploration of CXC receptors, and we found a decrease in receptor expression in severely injured patients but a concomitant increase in peroxide generation on a fibronectin surface in the vehicle control groups. Because we are still trying to tie CXC regulation to oxidant production, we have chosen to focus our study in the severely injured population. There is also an increase in the oxidant production in the moderately injured group, as defined by an ISS of greater than 9. Addressing the second question, potential interventions to decrease oxidant production in adherent neutrophils, I think that is very premature. The possible points of intervention are the integrins colocalizing with NADPH oxidase components or blockage of the cytoskeletal rearrangement induced by integrin signals that ultimately controls oxidant production. These potentially could be blocked with either a drug or antibody. The problem with interventions at this point is lack of understanding about what is occurring in vivo. Even in

676 Quaid et al

this study the neutrophils are explored in one or two dimensions at best, but in reality they are regulated and interacting in a four-dimensional system that contains a multitude of redundancy, so the likelihood of generating a meaningful intervention will not be possible until we better understand the biochemical interactions that occur both internally and externally to the cell. In regards to the third question about the variability of integrin expression and severity of injury and the variable results regarding anti-integrin antibodies, I believe the answer is found in this study. We show that the

Surgery October 2001 expression of the integrin is not the most important factor, but that integrin associations both internally and externally are critical to the regulation of specific functions. For example, we show that β1 specific matrix suggests a very complex signaling cascade regulating this event. Furthermore, neutrophils themselves contain a multitude of redundant pathways regulating a specific event such as oxidant production, so blockade of one pathway will not necessarily ensure complete inactivation of a specific function. This may also explain why antibodies are ineffective in vivo.