A chemotactic gradient sequestered on endothelial heparan sulfate induces directional intraluminal crawling of neutrophils

Abstract

During infection, chemokines sequestered on endothelium induce recruitment of circulating leukocytes into the tissue where they chemotax along chemokine gradients toward the afflicted site. The aim of this in vivo study was to determine whether a chemokine gradient was formed intravascularly and influenced intraluminal neutrophil crawling and transmigration. A chemokine gradient was induced by placing a macrophage inflammatory protein-2 (MIP-2)-containing (CXCL2) gel on the cremaster muscle of anesthetized wild-type mice or heparanase-overexpressing transgenic mice (hpa-tg) with truncated heparan sulfate (HS) side chains. Neutrophil-endothelial interactions were visualized by intravital microscopy and chemokine gradients detected by confocal microscopy. Localized extravascular chemokine release (MIP-2 gel) induced directed neutrophil crawling along a chemotactic gradient immobilized on the endothelium and accelerated their recruitment into the target tissue compared with homogeneous extravascular chemokine concentration (MIP-2 superfusion). Endothelial chemokine sequestration occurred exclusively in venules and was HS-dependent, and neutrophils in hpa-tg mice exhibited random crawling. Despite similar numbers of adherent neutrophils in hpa-tg and wild-type mice, the altered crawling in hpa-tg mice was translated into decreased number of emigrated neutrophils and ultimately decreased the ability to clear bacterial infections. In conclusion, an intravascular chemokine gradient sequestered by endothelial HS effectively directs crawling leukocytes toward transmigration loci close to the infection site.

Figure 1

Localized extravascular chemokine release induces directional intraluminal crawling of neutrophils toward infection site. (A) Number of adherent neutrophils per 100-μm venule and (B) number of emigrated neutrophils per field of view, in untreated (n = 3) WT cremaster muscle or activated by MIP-2 superfusion (n = 7) or an MIP-2-containing gel (n = 5). (C) Percentage of all adherent neutrophils that crawled within the observed vessel section before transmigration (MIP-2 SF, n = 67, 4 mice; MIP-2 gel, n = 100, 5 mice). (D-E) Displacement of crawling neutrophils (with crawling start and endpoints within the field of view) from the adhesion point (the center of the circle) to transmigration site (marked dots) during activation by (D) MIP-2 superfusion (n = 32, 4 mice) or by (E) an MIP-2-containing gel (n = 57, 5 mice) placed at 90 degrees at a distance of 400 μm from the venule. The vertical axis in the plot corresponds to the direction of blood flow (toward 0 degrees). *P < .05 compared with time 0 before addition of chemokine. †P < .05 vs WT untreated.

Figure 2

Localized extravascular chemokine release results in decreased crawling distance and crawling velocity. (A) Actual crawling distance and (B) calculated crawling velocity of neutrophils in response to MIP-2 superfusion (n = 32, 4 mice) or MIP-2–containing gel (n = 57, 5 mice). *P < .05 vs WT activated by MIP-2 superfusion.

Figure 3

Leukocyte transmigration is dependent on HS sequestration of chemokines on apical venular endothelium. (A) Number of adherent or (B) emigrated neutrophils, in the MIP-2-superfused cremaster muscle of WT (n = 7), hpa-tg mice (n = 7), or in WT pretreated with heparin (0.5 mg; n = 5). (C) In vivo intravascular binding of MIP-2 (expressed as FITC-MIP-2 intensity/ μm² vessel) 30 minutes after intra-arterial injection of MIP-2 conjugated to FITC in WT and hpa-tg cremasteric venules. *P < .05 compared with time 0 before addition of chemokine. †P < .05 vs WT at the same time period.

Figure 4

Chemokine gradients sequestered by endothelial HS effectively directs crawling leukocytes toward transmigration sites closer to the infection. (A) Displacement of crawling neutrophils (n = 48, 5 mice) from the adhesion point to transmigration site in hpa-tg cremaster muscle activated by an MIP-2-containing gel, placed extravascularly 400 μm from the observed venule (corresponding to 90 degrees in the polar chart). Representative confocal z-projections of (B) WT and (C) hpa-tg cremaster muscle activated with a FITC-MIP-2 (green) loaded gel. Endothelial cell junctions were stained with anti-CD31 mAb (red) and neutrophils with anti–Gr-1 mAb (blue) administered close intra-arterially.

Figure 5

Venular, but not arterial, endothelium sequesters MIP-2, and the chemokine is concentrated in junctional regions. Confocal microscopy images of cremasteric vessels (stained with anti-CD31 mAb). (A) FITC-conjugated MIP-2 (green) administered extravascularly in a gel is sequestered in and around venules (red, continuous arrow) but not in arterioles (red, dashed arrow). (B) MIP-2 administered intra-arterially is not sequestered in arterioles (green), but in (C) venules (green), as visualized by injection of fluorescently labeled anti–MIP-2 mAb (red). (D) Representative in vivo spinning disk confocal images of intravascular sequestration of FITC-labeled MIP-2 (green, administered in a gel) concentrated to endothelial cell junctions (red). Neutrophils were stained with anti–Gr-1 mAb (blue).

Figure 6

Random crawling in hpa-tg venules results in a decreased ability to clear bacterial infections. (A) Change in percentage of inoculated bacteria (bacterial bioluminescence) with time in WT (n = 9) and hpa-tg mice (n = 5), after subcutaneous administration of 10⁶ CFU of bioluminescent S aureus (strain Xen 29). *P < .05 vs WT. (B) Clearance rate of S aureus. (C) Representative images of S aureus bioluminescence detection in WT and hpa-tg mice at different time points after inoculation. Please note that the scales differ between time points but are the same at the same time point so that the mice can readily be compared.