Blockade of Class IA Phosphoinositide 3-Kinase in Neutrophils Prevents NADPH Oxidase Activation- and Adhesion-dependent Inflammation*

We examined the role of class IA phosphoinositide 3-kinase (PI3K) in the regulation of activation of NADPH oxidase in PMNs and the mechanism of PMN-dependent lung inflammation and microvessel injury induced by the pro-inflammatory cytokine TNF-α. TNF-α stimulation of PMNs resulted in superoxide production that was dependent on CD11b/CD18-mediated PMN adhesion. Additionally, TNF-α induced the association of CD11b/CD18 with the NADPH oxidase subunit Nox2 (gp91phox) and phosphorylation of p47phox, indicating the CD11b/CD18 dependence of NADPH oxidase activation. Transduction of wild-type PMNs with Δp85 protein, a dominant-negative form of the class IA PI3K regulatory subunit, p85α, fused to HIV-TAT (TAT-Δp85) prevented (i) CD11b/CD18-dependent PMN adhesion, (ii) interaction of CD11b/CD18 with Nox2 and phosphorylation of p47phox, and (iii) PMN oxidant production. Furthermore, studies in mice showed that i.v. infusion of TAT-Δp85 significantly reduced the recruitment of PMNs in lungs and increase in lung microvascular permeability induced by TNF-α. We conclude that class IA PI3K serves as a nodal point regulating CD11b/CD18-integrin-dependent PMN adhesion and activation of NADPH oxidase, and leads to oxidant production at sites of PMN adhesion, and the resultant lung microvascular injury in mice.

We examined the role of class IA phosphoinositide 3-kinase (PI3K) in the regulation of activation of NADPH oxidase in PMNs and the mechanism of PMN-dependent lung inflammation and microvessel injury induced by the pro-inflammatory cytokine TNF-␣. TNF-␣ stimulation of PMNs resulted in superoxide production that was dependent on CD11b/CD18-mediated PMN adhesion. Additionally, TNF-␣ induced the association of CD11b/CD18 with the NADPH oxidase subunit Nox2 (gp91 phox ) and phosphorylation of p47 phox , indicating the CD11b/CD18 dependence of NADPH oxidase activation. Transduction of wild-type PMNs with ⌬p85 protein, a dominant-negative form of the class IA PI3K regulatory subunit, p85␣, fused to HIV-TAT (TAT-⌬p85) prevented (i) CD11b/CD18-dependent PMN adhesion, (ii) interaction of CD11b/CD18 with Nox2 and phosphorylation of p47 phox , and (iii) PMN oxidant production. Furthermore, studies in mice showed that i.v. infusion of TAT-⌬p85 significantly reduced the recruitment of PMNs in lungs and increase in lung microvascular permeability induced by TNF-␣. We conclude that class IA PI3K serves as a nodal point regulating CD11b/CD18-integrin-dependent PMN adhesion and activation of NADPH oxidase, and leads to oxidant production at sites of PMN adhesion, and the resultant lung microvascular injury in mice.
PMNs 3 play an important antimicrobial role in protecting the host from a range of infectious agents by the generation of reactive oxygen species (ROS). Superoxide anion (O 2 . ) is formed from molecular oxygen (O 2 ) through the enzymatic activity of the NADPH oxidase complex (1). O 2 . gives rise to H 2 O 2 , spon-taneously or via superoxide dismutase, and to hypochlorous acid (HOCl) via myeloperoxidase, the PMN primary granule protein. However, excessive production of ROS by PMNs adherent to endothelial cells can also cause microvascular injury and tissue inflammation (2,3). The pro-inflammatory cytokine TNF-␣ is known to induce the production of ROS by surface-adherent PMNs (4). The dependence of TNF-␣-induced ROS production on PMN adherence to surfaces such as fibrinogen or endothelial cells involves the activation of ␤ 2 integrin, CD11b/CD18 (Mac-1, ␣ M ␤ 2 , CR3) (5,6). The molecular basis for the cross-talk between CD11b/CD18 and TNF-␣ receptors that regulate NADPH oxidase activation in PMNs is poorly understood. TNF-␣ interacts with two receptors, 60 and 80 kDa (7,8), both implicated in O 2 . generation by PMNs (9).
TNF-␣-activated signaling pathway in PMNs involves protein tyrosine phosphorylation, principally via tyrosine kinases of the Src family and Syk (10,11). However, TNF-␣ was shown to activate tyrosine phosphorylation only in adherent cells and required the interaction of TNF receptors with ␤ 2 integrins by as yet undefined mechanisms (7,11). Activation of phosphatidylinositol 3Ј-kinase (PI3K) induces the generation of the lipid second messenger, phosphatidylinositol 3,4,5-trisphosphate (PIP3) by phosphorylation of phosphatidylinositol 4,5-bisphosphate at the 3D position of the inositol ring (12). Class IA PI3K family members, which are activated by tyrosine kinases, consist of a heterodimer composed of a 110-kDa (p110 ␣, ␤, ␦) catalytic subunit and an adaptor/regulatory subunit (p85␣, p85␤, p55␣, p55␥, p50␣) (13,14). In contrast, Class IB enzymes contain the unique catalytic subunit (p110␥) associated with the regulatory subunit p101, and signal downstream of heterotrimeric G-protein-coupled receptors (15,16). Generation of PIP3 is essential for the activation of cellular responses such as O 2 . production in phagocytic cells (12,17) in response to multiple ligands, including TNF-␣ (10). However, the function of individual PI3K isoforms in PMNs has not been well established primarily because of the lack of isoform-selective and cell membrane-permeable reagents necessary for PMN studies. NADPH oxidase is an electron transport system that transfers reducing equivalents from NADPH to O 2 , resulting in the production of O 2 . (18 -22). NADPH oxidase in phagocytic cells consists of a membrane-integrated cytochrome b 558 , composed of two subunits (Nox2 (gp91 phox ) and p22 phox ) and three cytosolic proteins (p47 phox , p67 phox , and Rac2). NADPH oxidase exists in an unassembled state in resting cells, but upon stimulation, p47 phox , p67 phox , and Rac 2 translocate to the plasma membrane where they co-operatively form a complex with cytochrome b 558 ; however, the signaling mechanism regulating NADPH oxidase assembly in adherent PMNs and consequences of adhesion-dependent PMN activation in mediating vascular inflammation remain incompletely understood.
In the present study, we transduced PMNs with ⌬p85 protein, a dominant negative form (⌬478 -513 amino acids, p110 binding site) (23) of the class IA PI3K regulatory subunit, p85␣, which was fused to HIV-TAT (TAT-⌬p85) to facilitate cell membrane permeation of the protein. Our findings demonstrate that class IA PI3K activation is required for CD11b/CD18-dependent PMN adhesion, adhesion-dependent phosphorylation of p47 phox , and ROS production by PMNs. In studies in mice, i.v. infusion of TAT-⌬p85 protein prevented PMN-dependent lung microvascular injury and inflammation induced by TNF-␣. Thus, selective inhibitors of class IA PI3K represent a novel therapeutic strategy for treating lung inflammation and injury resulting from the inappropriate sequestration and adhesion-dependent activation of PMNs in microvessels.

EXPERIMENTAL PROCEDURES
Mice-C57BL/6 mice (WT) were obtained from Jackson Laboratories (Bar Harbor, ME). Mac-1-deficient mice (Mac-1 Ϫ/Ϫ ) were obtained from Dr. C. Ballantyne, Baylor College of Medicine, Houston, TX (24). We generated NIF transgenic mice (NIF ϩ/ϩ ) as described (25). All mice were housed under specific pathogen-free conditions with access to food and water ad libitum in the Animal Care Facility and all studies were made in accordance with institutional guidelines.
Isolation of Mouse PMNs-Neutrophils were purified from mouse bone marrow and peripheral venous blood using a discontinuous Percoll gradient as described with modifications (25,26). Purity of PMN preparations as assessed by examination of HEMA3 (Fisher)-stained cytospin (Shandon, Pittsburgh, PA) preparations (27) was ϳ90 -95% and viability assessed by Trypan blue exclusion was Ͼ95%.
Purification of TAT-⌬p85-A detailed method for the generation of dominant negative TAT-⌬p85 and TAT-GFP has been described (23,28,29). Briefly, a cDNA fragment encoding dominant negative p85 was amplified by PCR from the p85 cDNA with the deletion of 35 amino acids from residues 478 -513 with the insertion of two amino acids in pGEX. The PCR product was inserted into a pCRII TOPO vector. The plasmid was digested with AgeI/EcoRI and ligated into an AgeI/EcoRI-digested pTAT vector using T4 ligase. Purification of TAT fusion proteins was performed using modification of a method from Myou et al. (28). Briefly, TAT-⌬p85 was purified by sonication of high expressing BL21 Escherichia coli in 10 ml of buffer that was 20 mM HEPES, pH 8.0 and 100 mM NaCl. Cellular lysates were resolved by centrifugation, loaded onto 5-ml Ni-nitrilotriacetic acid column, washed, and eluted with 0.25-1.0 M imidazole in PBS. Imidazole was removed from the resultant protein solution using a PD-10 column. Each fusion protein preparation was flash-frozen at Ϫ80°C.
PMN Adhesion to Immobilized Fibrinogen-For PMN adhesion to immobilized fibrinogen, 96-well microplates were coated with 10 g/ml of fibrinogen overnight at 4°C and washed three times with HBSS before use. Mouse PMNs loaded with calcein-AM (2 g/ml) (Molecular Probes, Eugene, OR) for 30 min at 37°C were added to fibrinogen-coated plates. The assay for PMN adhesion was performed as described by us (25).
Microscopy of PMN Adhesion-Freshly isolated PMNs (2 ϫ 10 6 ), added to 8-well Lab-Tek II chamber slide with fibrinogencoated surface (10 g/ml, for 60 min at 37°C), were challenged with TNF-␣ at 37°C for 1 h. The cells were then fixed with 3.3% paraformaldehyde (Electron Microscopy Sciences) and washed. Photomicrographs of PMNs in contact with surfacecoated fibrinogen were acquired with a Zeiss LSM 510 confocal microscope.
PMN Superoxide Generation-PMNs without or with the treatment of TNF-␣ at 5 ϫ 10 6 cells/ml were seeded into a white, 96-well, flat-bottom tissue culture dish (E&K Scientific) coated with fibrinogen overnight. O 2 . generation by PMNs was measured as described, using isoluminol-enhanced chemiluminescence (30). Briefly, isoluminol was added to the cell suspension to a final concentration of 50 M, and horseradish peroxidase was added to a final concentration of 40 units/ml. O 2 .
Immunoprecipitation and Immunoblotting-Activation of Akt was assessed using a 1:500 dilution of an antibody that recognizes Akt phosphorylated on Ser 473 (Cell Signaling, Beverly, MA). An antibody against non-phosphorylated Akt (Cell Signaling, Beverly, MA) was used to assess protein loading.
Association of CD11b/CD18 with the NADPH oxidase subunit Nox2 was studied using a soluble enriched membrane fraction. After homogenization of the lungs in PBS (containing 1% protease inhibitors), the samples were centrifuged for 15 min at 14,000 ϫ g. The pellets were sonicated 2ϫ12 s in lysis buffer containing 100 mM Tris-HCl, pH 7.5, 1% Nonidet P-40, 5 mM EDTA, 5 mM EGTA, 150 mM NaCl, to which a protease inhibitor mixture (Sigma) was added. After centrifugation again for 15 min at 14,000 ϫ g, the supernatants of enriched membrane fraction were collected for immunoprecipitation. Equal amounts of protein (200 g/ml total protein) in supernatants were precleared using purified IgG followed by protein A-Sepharose beads. The precleared supernatants were recovered by centrifugation, and incubated overnight at 4°C with primary Abs (3 g ml Ϫ1 ) (antibody against CD11b, purchased from Santa Cruz Biotechnology, Santa Cruz, CA), followed by incubations with protein A-Sepharose beads for 2 h. The beads were collected by centrifugation (10,000 ϫ g, 5 min using an Eppendorf refrigerated microcentrifuge at 4°C), washed extensively (5 ϫ 10 min), solubilized and run on 4 -20% SDS-PAGE. After transfer to nitrocellulose, the membrane was then incubated with 1:500 anti-Nox2 (BD Biosciences) or anti-CD11b antibody overnight at 4°C. After washing, horseradish peroxidase-coupled secondary antibodies were then used to probe the membranes. The membranes were developed by manufacturer's instructions.
For analysis of total p47 phox phosphorylation, lungs and cells were homogenized in PBS containing 1% protease inhibitors, centrifuged at 14,000 ϫ g, and the pellet was sonicated in lysis buffer, as described above. Total cell lysates were immunoprecipitated with p47 phox antibody (from B. Babior and S. Catz, Scripps Research Institute, La Jolla, CA) and subjected to SDS-PAGE. Gels were stained with Pro-Q Diamond phosphoprotein gel stain (Molecular Probes, Eugene, OR) (32). The gels were scanned with Molecular Imager FX (Bio-Rad) and images of phosphoproteins were obtained. Non-phospho p47 phox loading of gels was obtained by Western blotting of total cell lysates.
TNF-␣ Challenge of Mice-Mice were challenged with TNF-␣ (100 g/kg, intraperitoneal). This dosage did not result in death within the 6-h experimental period after challenge. Control mice were injected intraperitoneal with an equal volume of PBS. Lungs were obtained at select time points after TNF-␣ challenge. One hour was used to assess the phosphorylation of Akt, p47 phox in lungs, leukocyte oxidative burst, lung PMN recruitment, and lung vascular permeability and edema formation. To determine the role of PI3K in acute lung injury, mice were first injected intraperitoneal with LY294002 (100 mg/kg of body weight, 100 l) or wortmannin (2 mg/kg of body weight, 100 l). In another group, animals receive i.v. TAT-⌬p85 (10 mg/kg) or a control injection via tail vein. The animals receiving injection of TAT-⌬p85 or control proteins (His-⌬p85 or TAT-GFP) were assigned randomly to the experimental groups consisting of six mice each. In preliminary studies, we demonstrated that TAT protein vector (administered as TAT-GFP) had no effect on TNF-␣-induced lung vascular permeability versus saline buffer control. The lungs obtained at different time points after TNF-␣ challenge were used to assess the changes as described above.
ROS Production by BAL Leukocytes-ROS in leukocytes obtained from mouse broncho-alveolar lavage (BAL) was measured as described (33). Briefly, leukocytes from BAL fluid were collected with or without intraperitoneal TNF-␣ challenge and incubated for 15 min with ROS-sensitive dye dichlorofluorescin-diacetate (DCFH-DA, 10 M, Molecular Probes) at 37°C in phosphate-buffered solution (PBS, Sigma Aldrich). After 1ϫ wash, the cells were suspended in PBS, and kept on ice until FACS analysis on the same day. The samples were analyzed by measuring fluorescence of 15,000 -20,000 events from the gated PMN population using a Coulter EPICS Elite ESP (Coulter Corporation, Miami, FL) with the laser at 530 nm. The fluorescent product of oxidation (DCFH) was measured by flow cytometry using four decade logarithmic amplification.
Lung PMN Sequestration-Sequestration of PMNs in lung tissue was assessed by determining myeloperoxidase (MPO) activity in lungs and by morphometrically quantifying PMN infiltration as described (2,27).
Pulmonary Microvascular Permeability and Edema Formation-K f,c was measured to determine pulmonary microvascular permeability to liquid and the rate of edema formation was continuously monitored by determining lung wet weight changes as described (25,27).
Statistical Analysis-Data are expressed as mean Ϯ S.E. Statistical analysis was performed using the 2-way analysis of variance and Newman-Keuls test for multiple comparisons. Significance was set at p Ͻ 0.05.

Effectiveness of TAT-⌬p85 in Preventing PI3K Activation
Caused by TNF-␣ in PMNs-We have shown that the synthesized TAT-⌬p85 fusion protein (useful for transducing human eosinophils (23)) specifically inhibits the activity of class IA PI3K. Experiments showed that all PMNs were transduced with TAT fusion protein after 2 min (data not shown). Western blot analysis showed that TAT-⌬p85 was also successfully transduced into mouse PMNs. TAT-⌬p85 expression was observed with little overlap with endogenous p85 after incubation of PMNs with 100 -300 nM TAT-⌬p85 (Fig. 1A, top panel). The functional efficacy of TAT-⌬p85 expression was assessed by its ability to inhibit phosphorylation of Akt, downstream target of PI3K. As shown in Fig. 1A (middle panel), the TNF-␣-induced phosphorylation of Akt was inhibited by TAT-⌬p85 in a concentration-dependent manner. The effectiveness of TAT-⌬p85 (300 nM) was similar to the nonspecific PI3K inhibitor LY294002 (50 M) used as a positive control (Fig.  1B). TAT-GFP (used as a control) had no effect on Akt phosphorylation (Fig. 1B).
Inhibition of Class IA PI3K Reduces PMN Adhesion to Immobilized Fibrinogen-We initially determined the role of the ␤ 2 integrin CD11b/ CD18 in mediating PMN firm adhesion to purified fibrinogen, a CD11b/CD18 ligand, induced by TNF-␣ (Fig. 2, A and B). PI3K inhibition by 300 nM TAT-⌬p85 or 50 M Ly294002 prevented PMN adhesion to fibrinogen induced by TNF-␣. Control peptide (TAT-GFP and TAT-p85) had no effect on PMN adhesion to fibrinogen induced by TNF-␣. PMN adhesion was also prevented by preincubation with anti-CD11b mAb ( Fig.  2A). Combination of TAT-⌬p85 and anti-CD11b mAb did not further reduce adhesion of PMNs to the plated fibrinogen ( Fig. 2A). To address definitively the role of CD11b/CD18 in mediating PMN firm adhesion in this system, PMNs isolated from mice either overexpressing the selective CD11b antagonist (34) NIF (NIF ϩ/ϩ ) or lacking CD11b (Mac-1 Ϫ/Ϫ ) were used for the adhesion assay. Compared with PMNs isolated from WT mice, we observed a decrease in adhesion to fibrinogen using the NIF ϩ/ϩ PMNs. Pretreatment of NIF ϩ/ϩ PMNs with TAT-⌬p85 did not further reduce PMN firm adhesion to fibrinogen (Fig. 2B). Similar results were obtained using Mac-1 Ϫ/Ϫ PMNs. As shown in Fig. 2C, TNF-␣ caused PMN spreading onto fibrinogencoated surface (Fig. 2C, middle) compared with untreated cells, which were round in appearance (Fig. 2C, left). The cells failed to spread when treated with TAT-⌬p85 prior to TNF-␣ challenge (Fig.  2C, right).
Inhibition of Class IA PI3K Reduces PMN Adhesion-dependent ROS Production-We also examined the role of CD11b/CD18 and  MARCH (Fig. 3).

PI3K Regulation of Neutrophil NADPH Oxidase
Because PMN activation and oxidant production are rapid signaling-dependent events (35), we also determined the early . release in a dose-dependant manner at early time points. Control peptide (TAT-GFP and TAT-p85) had no effect on TNF-␣-induced O 2 . release (Fig. 3B) Inhibition of Class IA PI3K Reduces CD11b/CD18 Association with Nox2 and p47 phox Activation-We examined whether NADPH oxidase subunits are co-localized with CD11b/ CD18 on PMN membrane after activation by TNF-␣. We found that CD11b/CD18 co-immunoprecipitated with Nox2 within 1 h of TNF-␣ challenge in PMNs. This response was attenuated by either pretreatment with anti-CD11b mAb or NIF (Fig. 4A).
We addressed the role of PI3K in signaling the interaction of CD11b/ CD18 with NADPH oxidase. As shown in Fig. 4B, pretreatment of WT PMNs with TAT-⌬p85 reduced the association of CD11b with Nox2. We also addressed whether PI3K regulates the TNF-␣-induced phosphorylation of p47 phox in PMNs. TNF-␣ induced p47 phox phosphorylation at 15-30 min, and introduction of TAT-⌬p85 prevented phosphorylation of p47 phox (Fig. 4C). Western blotting showed that the same amount of p47 phox was immunoprecipitated from each sample.
Class IA PI3K Regulates Activation of NADPH Oxidase in Vivo-We examined the efficacy of TAT-⌬p85 introduced in vivo in inhibiting PI3K activity in mouse lungs after systemic challenge with TNF-␣. Intravenous injection of 5-10 mg/kg TAT-⌬p85 or intraperitoneal injection of 100 mg/kg of LY294002 prevented the phosphorylation of Akt in lung homogenates (Fig. 5A). TAT-GFP had no inhibitory effect on phosphorylation of Akt. Systemic administration of TAT-⌬p85 also significantly reduced phosphorylation of p47 phox (Fig. 5B), whereas TAT-GFP had no inhibitory effect on the phosphorylation of p47 phox . . Inhibition of class IA PI3K and CD11b/CD18 ␤ 2 integrin reduces adhesion-dependent ROS production. A, PMNs (1 ϫ 10 5 ) isolated from WT, Mac-1 Ϫ/Ϫ , and NIF ϩ/ϩ mice were preincubated with indicated reagents for 15 min and then added to plated fibrinogen for 1 to 2 h. Superoxide production was measured as described. n ϭ 3. * denotes significant increase (p Ͻ 0.05) compared with unchallenged control. ** denotes significant decrease (p Ͻ 0.05) compared with TNF-␣ alone-treated group. † denotes significant decrease (p Ͻ 0.05) in PMNs isolated from genetically modified mice compared with PMNs isolated from WT mice post-TNF-␣ stimulation. Bars indicate mean Ϯ S.E. B, superoxide generation was determined in real-time based on isoluminol-ECL. cps, counts per second of light emitted. ϳ0.2 ϫ 10 6 cells/sample were preincubated with either 100 nM, 300 nM TAT-⌬p85, or 300 nM TAT-GFP for 15 min prior to stimulation with TNF-␣ (40 ng/ml). The figure shows representative curves from one of three independent experiments that produced similar results.
Because phosphorylation of p47 phox is a key determinant of NADPH oxidase activation, we surmised that CD11b/CD18 activation was involved in the response. We observed a reduction in phosphorylation of p47 phox in TNF-␣-challenged lungs of Mac-1 Ϫ/Ϫ mice; similar results were also obtained with WT mice pretreated with the anti-CD11b mAb prior to TNF-␣ challenge (Fig. 5C).
PI3K Regulates TNF-␣-dependent ROS Production in BAL Leukocytes-ROS production was measured in BAL leukocytes obtained from TNF-␣-challenged WT mice at 1 h compared with buffer-challenged control mice (Fig. 6A); the response persisted up to 6 h (data not shown). To address the in vivo relevance of class IA PI3K in TNF-␣-induced oxidant production in leukocytes, mice were pretreated with the peptide inhibitor of PI3K activity described above. In positive control experiments, the PI3K inhibitors wortmannin or LY 294002 significantly reduced oxidant generation in leukocytes obtained from BAL of TNF-␣-challenged mice (Fig. 6, A and B). Systemic pretreatment with TAT-⌬p85 in a concentration-dependent manner reduced ROS production in the BAL leukocytes (Fig. 6C). Pretreatment of WT mice with anti-CD11b mAb also significantly reduced ROS production after TNF-␣ challenge (Fig. 6F). How-ever, no additive reduction of ROS production was seen with combination of both TAT-⌬p85 and anti-CD11b mAb (Fig.  6D). In contrast, TAT-GFP had no effect on oxidant production by the BAL leukocytes (Fig. 6E).
Class IA PI3K Mediates TNF-␣-induced PMN Sequestration and Vascular Injury in Mouse Lungs-To address the in vivo role of PI3K in mediating the TNF-␣-induced PMN recruit-  A, efficacy of TAT-⌬p85 in inhibiting PI3K activity in mouse lung after TNF-␣ challenge. Mice were pretreated with buffer or TAT-⌬p85 (5-10 mg/kg, i.v.), TAT-GFP (10 mg/kg, i.v.), or LY294002 (100 mg/kg, intraperitoneal) for 30 min and then challenged with TNF-␣ (100 g/kg, intraperitoneal). Akt phosphorylation in lung tissue at 1 h after TNF-␣ challenge was measured by Ser 473 phosphorylationspecific antibody. Total levels of Akt were measured as a loading control. Results are representative of three independent experiments. B, effect of PI3K inhibition on phosphorylation of p47 phox in lung tissues caused by TNF-␣ challenge. Mice were treated as above, and total cell lysates were immunoprecipitated with p47 phox antibody. P47 phox phosphorylation was measured by fluorescent staining of SDS-polyacrylamide gels using Pro-Q Diamond phosphoprotein gel stain kit as described under "Experimental Procedures." Results are representative of three independent experiments. C, p47 phox phosphorylation in WT lungs versus Mac-1-deficient lungs as well as the effect of CD11b blockade on phosphorylation of p47 phox in lung tissues induced by TNF-␣. p47 phox phosphorylation was measured as above. Equal amounts of protein (500 g of lung lysate) were added to the immunoprecipitation (IP) reaction. Results are representative of three independent experiments. MARCH 2, 2007 • VOLUME 282 • NUMBER 9 ment, PMN number in lungs was measured as described (27). Lung tissue PMN counts increased from 551.3 Ϯ 52.1 cells/30 mm 2 lung tissue to 3395.3 Ϯ 73.8 cells/30 mm 2 1 h after TNF-␣ challenge in WT mice. Systemic pretreatment with TAT-⌬p85 reduced lung PMN sequestration after TNF-␣ challenge (Fig.  7). PMN sequestration was reduced by ϳ55% at 1 h by TAT-⌬p85, whereas the control p85 protein had no effect (Fig. 7A). TNF-␣-induced PMN sequestration was reduced ϳ40% by anti-CD11b mAb (Fig. 7A) whereas isotype-matched control antibody IgG2a did not affect PMN sequestration in lungs. No additive blockade of PMN sequestration was observed after combination of both anti-CD11b and TAT-⌬p85. PMN sequestration was also significantly reduced in Mac-1 Ϫ/Ϫ and in NIF ϩ/ϩ mice (Fig. 7A). Systemic pretreatment with TAT-⌬p85 also reduced lung myeloperoxidase (MPO) activity by ϳ51% at 1 h after TNF-␣ challenge (Fig. 7B), whereas the control p85 protein had no effect.

PI3K Regulation of Neutrophil NADPH Oxidase
Activation of PMN adherent to the microvessels plays a key role in lung inflammation and contributes to the mechanism of lung microvascular injury by the release of oxidants, proteases, chemokines, and other mediators (25,37), because the above data show an important role for class IA PI3K in signaling PMN ROS generation and sequestration in lungs following TNF-␣ challenge, we addressed the possibility that TAT-⌬p85 would also prevent lung vascular microvascular injury induced by TNF-␣. Pulmonary capillary filtration coefficient (K f,c ), a measure of microvascular permeability, and lung edema formation were determined in WT lungs perfused with TNF-␣. TNF-␣ challenge resulted in increased K f,c (Fig. 8A). K f,c value at 1 h after TNF-␣ was elevated ϳ6-fold over basal, indicating a marked increase in microvessel permeability. Systemic treatment with TAT-⌬p85, but not TAT-GFP, prevented the TNF-␣-induced increase in lung microvascular permeability and edema formation (Fig. 8A). In negative control experiments, the increases in K f,c induced by TNF-␣ were reduced by ϳ50% in Mac-1 Ϫ/Ϫ mice and ϳ75% in NIF ϩ/ϩ mice (Fig. 8A) and by pretreatment with either NIF or anti-CD11b mAb (Fig.  8A). No further reduction in K f.c was observed after combination of both anti-CD11b mAb and TAT-⌬p85 in WT mice. TAT-⌬p85 did not prevent further the lung vascular permeability response seen in Mac-1 Ϫ/Ϫ or NIF ϩ/ϩ mice after TNF-␣. Fig. 8B shows the protective effect of TAT-⌬p85 on lung edema formation. TAT-⌬p85, but not TAT-GFP reduced lung wet weight increase after TNF-␣ challenge (Fig. 8B) and no further protective effect after combined treatment with anti-CD11b and TAT-⌬p85. Fig.  8C shows the protective effect of anti-CD11b mAb on lung edema formation, whereas anti-CD11a mAb was not protective. TAT-⌬p85 had no further inhibitory effect on lung wet weight increases seen in Mac-1 Ϫ/Ϫ (Fig. 8D) or NIF ϩ/ϩ (Fig. 8E) mice.

DISCUSSION
The augmented oxidant generation by the adherent PMNs may be a crucial factor mediating lung inflammation and microvascular injury (25,37,38). Because oxidant generation is linked to ␤ 2 integrin-dependent PMN adhesion, we addressed the role of the PMN ␤ 2 integrin CD11b/CD18 (Mac-1) in activating PMN NADPH oxidase complex, and thereby in the mechanism of PMN-mediated lung inflammation and microvascular injury. We showed that (i) activation of the class IA PI3K isoform is an important signal mediating PMN adhesion

PI3K Regulation of Neutrophil NADPH Oxidase
and adhesion-dependent ROS production by PMNs, (ii) class IA PI3K isoform mediates this response by inducing CD11b/ CD18 interaction with Nox2 subunit of NADPH oxidase, and (iii) selective inhibition of class IA PI3K prevents lung PMN sequestration and microvascular injury induced by TNF-␣ challenge of mice.
Previous studies have shown involvement of PI3K in signaling the activation of NADPH oxidase by G protein-coupled receptors (GPCRs) such as the fMLP receptor (39) and certain non-GPCRs such as TNF receptors (40). However, the isoforms of PI3K regulating NADPH oxidase activation in PMNs have not been clearly delineated, in part, because of the lack of iso-form-specific inhibitors. Prior studies in PI3K␥-deficient mice showed impaired fMLP-stimulated PMN respiratory burst activity (17,41) indicating the important role of GPCR-coupled PI3K␥ in the activation of NADPH oxidase. Using dominant negative protein inhibitors of class IA and IB PI3K, a recent study has demonstrated the dominant role of class IA PI3K in fMLPstimulated superoxide generation (36). In the present study, we addressed the role of class IA PI3K in mediating TNF-␣-induced activation of PMN NADPH oxidase and its consequences in the mechanism of lung inflammation and injury. Studies were made using the ⌬p85 protein, a dominant negative form of the class IA PI3K regulatory subunit, p85␣, fused to HIV-TAT (TAT-⌬p85) enabling the protein to readily permeate the plasmalemmal barrier of PMNs (32). We observed that PI3K inhibition by TAT-⌬p85 significantly reduced CD11b/CD18dependent PMN adhesion to fibrinogen and ROS generation induced by TNF-␣ challenge. TAT-⌬p85 also abrogated the phosphorylation of p47 phox and association of CD11b/CD18 with NADPH oxidase. In mouse studies, we showed that TAT-⌬p85 i.v. infusion significantly reduced the recruitment of PMNs into lungs and the increase in lung microvascular permeability induced by TNF-␣. These results demonstrate the critical role of class IA PI3K-activated NADPH oxidase and resultant ROS production in the mechanism of lung inflammation and microvascular injury.