The role of endothelial PI3Kgamma activity in neutrophil trafficking

Abstract

Phosphoinositide 3-kinase gamma (PI3Kgamma) in neutrophils plays a critical role in the directed migration of these cells into inflamed tissues. In this study, we demonstrate the importance of the endothelial component of PI3Kgamma activity relative to its leukocyte counterpart in supporting neutrophil interactions with the inflamed vessel wall. Despite the reconstitution of class-Ib PI3K function in neutrophils of p110gamma-/- mice, we observed a 45% reduction in accumulation of these cells in an acute lung injury model. Mechanistically, this appears to result from a perturbation in selectin-mediated adhesion as manifested by a 70% reduction in wild-type (WT) neutrophil attachment to and 17-fold increase in rolling velocities on p110gamma-/- microvessels in vivo in response to tumor necrosis factor alpha (TNFalpha). This alteration in adhesion was further augmented by a deficiency in p110delta, suggesting that the activity of both catalytic subunits is required for efficient capture of neutrophils by cytokine-stimulated endothelium. Interestingly, E-selectin-mediated adhesion in p110gamma-/-) mice was impaired by more than 95%, but no defect in nuclear factor kappa B (NF-kappaB)-induced gene expression was observed. These findings suggest a previously unrecognized partnership between class-I PI3Ks expressed in leukocytes and endothelium, the combination of which is required for the efficient trafficking of immunocompetent cells to sites of inflammation.

Figure 1.

Contribution of the nonleukocyte component of class-Ib PI3K activity to neutrophil accumulation in response to LPS. GFP-expressing cell counts per unit area obtained 6 hours after intratracheal administration of LPS (10 μg/g) in mice (A) deficient in or chimeric for PI3Kγ activity, (B) deficient in or chimeric for PI3Kδ activity, (C) pretreated with the function-blocking or nonfunction-blocking F(ab')₂ 9A9 and CL37, respectively. (A) ▪ indicates WT littermate (GFP^–/+ p110γ^+/+); ▪, GFP^–/+ p110γ^–/–; and , p110γ^–/– reconstituted with GFP^–/+ p110γ^+/+ fetal liver cells. (B) ▪ indicates WT littermate (GFP^–/+ p110δ^+/+); ▪, GFP^–/+ p110δ^–/–; and , p110δ^–/– reconstituted with GFP^–/+ p110δ^+/+ fetal liver cells. (C) ▪ indicates WT littermate (GFP^–/+ p110δ^+/+); ▪, WT littermate plus mAB CL37; and , WT littermate plus mAb 9A9. Mean values (± SD) are shown for 8 mice in each experimental or control group. *P < .05.

Figure 2.

Contribution of p110γ to TNFα production. LPS-induced generation of TNFα in whole blood harvested from either 3 p110γ^–/– or WT-matched controls. Mean values (± SE) for experiments performed in triplicate.

Figure 3.

p110γ is functionally expressed in venular endothelium. (A) Immunoblots of the p110γ catalytic subunit using recombinant proteins (i) and lysates obtained from HUVECs or neutrophils (ii), or the murine endothelioma cell line bEND3.1 (iii). Measurement of (B) PI3Kγ and (C) PI3Kδ kinase activity. Immunoprecipitates of p110γ and p110δ (▪) from HUVEC lysates were assayed for PI3K activity with or without the addition of 10 μM IC87114 or LY294002 as described in “Materials and methods.” ▪ indicates control. The results are expressed as the percent of activity in untreated immunoprecipitates and represent the mean plus or minus SD values of 3 independent determinations in duplicate. *P < .01 as compared with untreated cells.

Figure 4.

A deficiency in endothelial p110γ impairs neutrophil accumulation on TNFα-stimulated CM venules. Rolling fraction (the number of GFP-expressing cells that attached and rolled on inflamed endothelium divided by the total flux of cells passing through the vessel) determined in mice (A) deficient in or chimeric for p110γ, (B) deficient in or chimeric for p110δ activity. ▪ indicates WT littermate (WT GFP^–/+); ▪, GFP^–/+ p110γ^–/– or GFP^–/+ p110δ^–/–; and , p110γ^–/– or p110δ^–/– reconstituted with WT GFP^–/+ fetal liver cells. Results represent the mean plus or minus SD; *P < .05 as compared with WT littermates. n = number of mice/venules analyzed. Rolling velocities for consecutive interacting cells (n = 30 per venule) in mice (C) deficient in or chimeric for p110γ, or (D) deficient in or chimeric for p110δ. Data represent the mean ± SD for more than 150 cells per experimental condition; *P < .05. (E) Rolling fraction and (G) rolling velocities for p110γ^–/–/p110δ^–/– mice reconstituted with FLC from GFP^–/+ WT fetal liver cells (▪). ▪ indicates WT littermates reconstituted with GFP^–/+ WT fetal liver cells. *P < .05 as compared with controls. (F) Representative intravital micrographs depicting the extent of neutrophil adhesion to and transmigration across CM venules in WT mice (i) and p110γ^–/– (ii) or p110δ^–/– (iii) animals reconstituted with WT FLCs (3 hours after stimulation with TNFα).

Figure 5.

The p110γ is required for efficient capture of neutrophils by E-selectin. (A) Immunoprecipitates of spleens from WT mice using a p110δ-specific antibody that was assayed for PIP₃ production in the absence (▪) or presence (▪)of IC87114 (10 μM). The results are expressed as the percent activity in vehicle-treated p110δ immunoprecipitates and represent the mean plus or minus SD values of 3 independent determinations in duplicate. *P < .01 as compared with vehicle-treated WT spleen lysate. Rolling fraction of P-selectin^–/–/GFP^–/+ animals (B) pretreated with vehicle control (▪), IC87114 (▪), or mAb 9A9 () 1 hour prior to induction of inflammation with TNFα or (C) in those that also lacked p110γ. Data represent the mean plus or minus SD and are normalized as a percentage of control; *P < .05 as compared with control. n = number of mice /venules analyzed. (D) In vivo imaging of E-selectin expression on CM venules in mice with or without prior stimulation with TNFα. Fluorescent Qdots coated with an antibody that recognizes murine E-selectin were injected intravenously into animals and immunolocalization of this adhesion molecule was visualized by intravital microscopy. Transmitted light (i) and epifluorescence (ii) images depict staining of venules in P-selectin^–/– mice in the absence of TNFα stimulation. Representative micrographs of inflamed venules in P-selectin^–/– animals pretreated with (iii) IC87114 or (iv) lacking p110γ, respectively, or mice deficient in E-selectin in the absence (v) or presence (vi) of IC87114.

Figure 6.

Role of class-I PI3Ks in E-selectin expression in vitro. (A) E-selectin and PECAM-1 surface expression on HUVECs (± 5 μM TNFα) pretreated (1 hour) with the p110δ-specific inhibitor IC87114 (2 μM), the pan–class-I PI3K inhibitor LY294002 (10 μM), or the NF-κB function blocker Bay 11-7082 (10 μM). Adhesion molecule expression on nonstimulated cells is shown for comparison (top panels). Black lines represent vehicle control or cells treated with the indicated inhibitors (gray). (B) E-selectin expression on TNFα-stimulated HUVECs remains unchanged even at doses of IC87114 known to inhibit p110γ (50 μM). (C) TNFα-mediated NF-κB translocation in nuclear extracts from HUVECs pretreated with vehicle control or IC87114 as quantified by ELISA for the p50 subunit. Data represent the mean plus or minus SD for 1 of 3 experiments with similar results. *P < .01 as compared with cells treated with TNFα alone. (D) Percent reduction in attachment of purified human neutrophils to HUVEC monolayer that was pretreated with the indicated inhibitors prior to stimulation with TNFα (4 hours). Results are expressed as the percentage of cells that bound to the vehicle-treated substrate at a wall shear rate of 200 s^–1. Data represent the mean plus or minus SD for 3 experiments performed in duplicate. *P < .05 as compared with vehicle control.

Figure 7.

Role of class-I PI3Ks in neutrophils in adhesion in vitro. Attachment of (A) p110γ^–/– versus (B) p110δ^–/– neutrophils to surface-immobilized E- or P-selectin Ig chimeras at a wall shear rate of 200 s^–1. Results are compared with neutrophils purified from WT littermates. Adhesion to the β₂-integrin ligand ICAM-1-Ig chimera is shown as control. Results represent the mean plus or minus SD for 3 experiments performed in duplicate.