Spleen tyrosine kinase Syk is necessary for E-selectin-induced alpha(L)beta(2) integrin-mediated rolling on intercellular adhesion molecule-1

Abstract

Engagement of neutrophils by E-selectin results in integrin activation. Here, we investigated primary mouse neutrophils in whole blood by using intravital microscopy and autoperfused flow chambers. Slow rolling on E-selectin coimmobilized with intercellular adhesion molecule-1 (ICAM-1) required P-selectin glycoprotein ligand (PSGL)-1, was dependent on alpha(L)beta(2) integrin (LFA-1), and required continuous E-selectin engagement. Slow rolling was abolished by pharmacological blockade of spleen tyrosine kinase (Syk) and was absent in Syk(-/-) bone-marrow chimeric mice. Treatment with tumor necrosis factor-alpha lowered rolling velocity further and induced CXC chemokine ligand-1 (CXCL1) and CXC chemokine receptor-2 (CXCR2)-dependent leukocyte arrest on E-selectin and ICAM-1. Arrest but not rolling was blocked by an allosteric inhibitor of LFA-1 activation. Neutrophil recruitment in a thioglycollate-induced peritonitis model was almost completely inhibited in Selplg(-/-) mice or Syk(-/-) bone-marrow chimeras treated with pertussis toxin. This identifies a second neutrophil-activation pathway that is as important as activation through G protein-coupled receptors (GPCRs).

Figure 1. PSGL-1, but not CD44, is required for neutrophil slow rolling on ICAM-1 upon E-selectin engagement

Carotid cannulas were placed in untreated WT and CD44^-/- mice and connected to E-selectin and E-selectin+ICAM-1-coated autoperfused flow chambers. Rolling velocity of neutrophils presented as mean ± SEM (A). Rolling velocity of neutrophils from untreated WT and PSGL-1 deficient mice on E-selectin and E-selectin+ICAM presented as mean ± SEM (B). Rolling velocity of untreated neutrophils on E-selectin+ICAM-1 before and after blocking LFA-1 or Mac-1 or both (C). (D) Rolling velocity of neutrophils from untreated mice on P-selectin and P-selectin+ICAM-1. (E) Analysis of the rolling velocity of neutrophils in inflamed cremaster venules of P-selectin^-/- mice 2h after TNF-α injection. (F) Neutrophil adhesion in the flow chamber after co-immobilization of CXCL1 (10 µg/ml) with P-selectin+ICAM-1 or E-selectin+ICAM-1. The wall shear stress in all flow chamber experiments was 5.94 dynes/cm². At least three mice and four flow chambers per group. (G-I) Intravital microscopy of the cremaster muscle from P-selectin deficient mice and E-selectin deficient mice 2h after TNF-α application. Rolling velocity (G), rolling flux fraction (H), and number of adherent cells (I).At least four mice and 20 vessels per group. * P < 0.05.

Figure 2. Continuous E-selectin engagement is required to keep LFA-1 in an activated conformation

(A) The first half of autoperfused flow chambers was coated with E-selectin+ICAM-1 and the second half with ICAM-1 alone. Cells rolled on E-selectin+ICAM-1 but detached immediately upon reaching ICAM-1 alone (representative of four independent experiments). (B) Number of rolling neutrophils on E-selectin+ICAM-1 coated chambers before and after injection of a blocking E-selectin antibody (9A9). Data presented are the mean ± SEM from at least four mice and four flow chambers (n=4). ^# P < 0.05.

Figure 3. Syk is involved in integrin activation after E-selectin or P-selectin binding to PSGL-1

Whole blood was perfused through E-selectin and E-selectin+ICAM-1 coated autoperfused flow chambers at the indicated wall shear stress. The rolling velocity of neutrophils from mice pre-treated with piceatannol (1mg/mouse, A), a specific Syk inhibitor, and Syk^-/- chimeric mice (B) were determined after 6 minutes (n=4). (C) Rolling velocity on E-selectin and E-selectin+ICAM-1 was measured after blocking of the mitogen-activated protein (MAP)-kinase p38 by the inhibitor SB203580 (100 μg/mouse i.p.) 1h prior to the experiment (n=3). ^# P < 0.05.

Figure 4. TNF-α, but not IL-1β, causes neutrophil arrest on E-selectin + ICAM-1

(A) Carotid cannulas of TNF-α or IL-1β pre-treated mice were connected to E-selectin and E-selectin+ICAM-1-coated autoperfused flow chambers. Rolling velocity of neutrophils was determined at a wall shear stress of 5.94 dyn/cm² (n=4) (B, C) Number of adherent neutrophils on E-selectin+ICAM-1 of TNF-α pre-treated mice or whole blood after 6 minutes. Mice were pre-treated with TNF-α and an allosteric inhibitor (Compound 6845) of LFA-1 or vehicle and rolling velocity (D) and number of adherent neutrophils (E) were determined on E-selectin and E-selectin+ICAM-1. To ensure that only LFA-1-dependent rolling and adhesion were investigated, all mice were treated with a blocking monoclonal antibody to Mac-1 (n=4). * and ^# P < 0.05.

Figure 5. CXCR2 and Gαi, but not Syk, control neutrophil arrest in response to TNF-α

Rolling velocity of neutrophils from TNF-α pre-treated mice on E-selectin and E-selectin+ICAM-1 coated flow chambers after blocking Gαi with PTx or Syk by piceatannol alone or in combination (A). (B) Neutrophil adhesion on E-selectin+ICAM-1 coated flow chambers after blocking Gαi with PTx or Syk by piceatannol alone or in combination of TNFα pre-treated mice. Rolling velocity (C) and neutrophil adhesion (D) on E-selectin+ICAM-1 coated flow chambers after blockade of CXCR2 by antibody or of Syk by piceatannol alone or in combination following TNF-α application. Cremaster muscles of WT mice were stained for CXCL1 without (E) and with (F) TNF-α. TNF-α injection induced an increase in adherent leukocytes in the microcirculation of the cremaster muscle in WT mice compared to untreated mice. Inserts show typical neutrophils stained for CXCL1 (E, F). (G) CXCL1 concentration in plasma before and 2h after TNF-α injection (n=4). ^# P < 0.05.

Figure 6. Model of LFA-1 activation during neutrophil rolling

E-selectin binding to PSGL-1 leads, through Syk (arrows), to partial LFA-1 (open curved arrow) activation that supports rolling on ICAM-1 (shown as dimer). CXCR2 engagement by CXCL1 leads, through Gai, to full LFA-1 activation, resulting in arrest on ICAM-1 (filled curved arrow). Straight arrows indicate signaling pathway, but interactions may be indirect. LFA-1 conformations from Takagi et al (Takagi and Springer, 2002).

Figure 7. Inhibition of Gαi in PSGL-1^-/- mice blocks neutrophil recruitment in vivo

Peritoneal neutrophil influx 4 h after 1 ml injection of 4% thioglycollate into WT mice, WT mice that had received 4 μg PTx before thioglycollate injection, PSGL-1^-/- mice, PSGL-1^-/- mice plus PTx, E-selectin^-/- mice, E-selectin^-/- mice plus PTx, WT mice that had received 1mg piceatannol (picea) and 4 μg PTx, and Syk^-/- chimeric mice (Syk^-/-) plus 4 μg PTx (n=5 each). Data expressed as total numbers of neutrophils in the peritoneal lavage fluid counted using Kimura-stained samples, ^# P < 0.05.