Selectin catch-bonds mechanotransduce integrin activation and neutrophil arrest on inflamed endothelium under shear flow - PubMed (original) (raw)

Clinical Trial

Selectin catch-bonds mechanotransduce integrin activation and neutrophil arrest on inflamed endothelium under shear flow

Vasilios A Morikis et al. Blood. 2017.

Abstract

E-selectin extends from the plasma membrane of inflamed endothelium and serves to capture leukocytes from flowing blood via long-lived catch-bonds that support slow leukocyte rolling under shear stress. Its ligands are glycosylated with the tetrasaccharide sialyl Lewisx (sLex), which contributes to bond affinity and specificity. E-selectin-mediated rolling transmits signals into neutrophils that trigger activation of high-affinity β2-integrins necessary for transition to shear-resistant adhesion and transendothelial migration. Rivipansel is a glycomimetic drug that inhibits E-selectin-mediated vaso-occlusion induced by integrin-dependent sickle-red blood cell-leukocyte adhesion. How Rivipansel antagonizes ligand recognition by E-selectin and blocks outside-in signaling of integrin-mediated neutrophil arrest while maintaining rolling immune-surveillance is unknown. Here, we demonstrate that sLex expressed on human L-selectin is preferentially bound by E-selectin and, on ligation, initiates secretion of MRP8/14 that binds TLR4 to elicit the extension of β2-integrin to an intermediate affinity state. Neutrophil rolling over E-selectin at precise shear stress transmits tension and catch-bond formation with L-selectin via sLex, resulting in focal clusters that deliver a distinct signal to upshift β2-integrins to a high-affinity state. Rivipansel effectively blocked formation of selectin catch-bonds, revealing a novel mechanotransduction circuit that rapidly converts extended β2-integrins to high-affinity shear-resistant bond clusters with intracellular adhesion molecule 1 on inflamed endothelium.

PubMed Disclaimer

Figures

None

Graphical abstract

Figure 1.

Figure 1.

PMN arrest and transmigration on inflamed endothelium is inhibited by Rivipansel. Isolated human PMNs were perfused over IL-1β-stimulated human umbilical vein endothelial cell monolayers in a microfluidic flow chamber at physiological shear stress of 2 dynes/cm2. (A) PMN rolling to arrest and transmigration across human umbilical vein endothelial cell monolayers are superimposed. PMNs are phase bright until migration below endothelial monolayer, then become phase dark (marked by white arrows; scale bar, 10 μm). (B) PMN density, adherent on inflamed human umbilical vein endothelial cells measured after 2 minutes of shear flow. Arrest and TEM averaged from 5 fields of view represents mean ± SEM for 40 cells (n = 3 separate experiments). Significant differences are indicated for PMN arrest and TEM compared with stimulation and no shear condition. ***P < .001; **P < .01; *P < .05. (C) PMN rolling at 3 minutes, arrest, and TEM density on human umbilical vein endothelial cells at 7 min in the presence and absence of Rivipansel inhibition are superimposed. Data points represent mean ± SEM from 5 fields per condition (n = 4 separate experiments). Rolling (*), arrest ($), and TEM (#) significance (at P < .01) compared with untreated reported. (D) PMN arrest efficiency (number of PMNs captured/count in blood) of whole blood sheared over recombinant E-selectin/ICAM-1 for 4 patients treated for sickle cell disease before (t = 0) and after Rivipansel infusion (left _y_-axis) and average serum concentrations of Rivipansel after infusion (right _y_-axis). Each data point is the mean ± SEM of 5 to 11 fields of view along the centerline of the channel. ***P < .001; **P < .005; *P < .01 denote significance from baseline at t = 0). TEM, transendothelial migration.

Figure 2.

Figure 2.

E-selectin binds sLe x on L-selectin and inhibition with Rivipansel. Isolated human PMNs were treated with IgG, anti-PSGL-1 antibody (KPL-1), anti-L-selectin (DREG56), the combination, or anti-β2-integrin (IB4) antibody and perfused over (A) E/I or (B) P/I substrate in a microfluidic chamber at 2 dynes/cm2. PMN rolling (#P < .05; ##P < .001) and arrest ($P < .01) significance compared with IgG control, subtracting background adhesion. Values depicted are mean ± SEM for n = 3 separate experiments. L-selectin and PSGL-1 were immunoprecipitated from isolated human PMNs with recombinant E-selectin or P-selectin treated with vehicle, or in the presence of Rivipansel (100 μM), or GSnP-6 (120 μM). Western blot protein content of (C) E-selectin-IgG pulldown and quantitation of the ratio of (D) L-selectin (∼69 kDa) or (E) PSGL-1 (∼120 kDa + ∼210 kDa) to untreated immunoprecipitation of total protein. (F) P-selectin pulldown of (G) L-selectin and (H) PSGL-1 relative to total protein pulldown. Relative density of L-selectin and PSGL-1, respectively, was compared between mean ± SEM, as depicted (n = 3 separate experiments). *P < .05; **P < .01; ***P < .001.

Figure 3.

Figure 3.

sLe x glycomimetics alter rolling to arrest on E-selectin and bond mechanics. (A) PMNs rolling to arrest over E/I substrates (<0.4 μm/sec) treated with vehicle (PBS), Rivipansel, or GSnP-6 was measured and normalized to total number of interacting PMNs; data are reported as mean ± SEM (n = 3 separate experiments; *P < .05; ***P < .001 compared with vehicle). (B) PMN rolling velocity was quantified over an E-selectin substrate treated with vehicle, Rivipansel (6.5 μM), or GSnP-6 (12 μM) and binned at intervals of 0.4 μm/sec. Comparison of GMI-1070 showed *significance and GSnP-6 showed ***significance over untreated controls (n = 3 separate experiments; *P < .05 and ***P < .001). (C) Schematic depicts dynamic interaction between PMNs and recombinant E-selectin-coated protein-G beads recorded at 50 frames per second in the flow channel. Isolated human PMNs were treated with an anti-Mac-1 and anti-PSGL-1 blocking antibodies and then perfused through flow chambers. PMNs pivot over beads and pull a membrane tether at defined wall shear stress. Adhesive interactions were identified as collisions that had a visible pause in PMN motion for at least 1 frame, along with velocities below the hydrodynamic velocity. (D) Tether duration was compared with step-wise increases in calculated tether forces. (E) Tether efficiency (collisions resulting in adhesion divided by the total collisions observed) as shear stress was ramped in a stepwise manner. Data were reported as mean ± SEM (n = 3 separate experiments; **P < .01; *P < .05 glycomimetics compared with 9 pN tether force and 0.25 dynes/cm2 wall shear stress).

Figure 4.

Figure 4.

Glycomimetic antagonism of selectin mediated β 2 -integrin activation. (A) PMN rolling on a substrate of E-selectin in the presence of vehicle control, Rivipansel (6.5 μM), or GSnP-6 (12 μM) was dynamically imaged using qDF to detect L-selectin (AF488 anti-human DREG55) and PSGL-1 (PE anti-human PL-1) engagement in the plane of adhesive contact. (B) L-selectin (FITC DREG-55) and (C) PSGL-1 (PE PL-2) receptor cluster area and frequency were determined and reported as mean ± SEM (n = 3 separate experiments; **P < .01 compared with vehicle). (D) PMNs imaged by TIRF were pretreated with a membrane dye (DiI) and HA β2-integrin reporter antibody (mAb24) and perfused over E/I substrates treated with vehicle, Rivipansel (6.5 μM), or GSnP-6 (12 μM). (E) HA β2-integrin cluster number (MFI > 10 pixels above background more than ∼0.1 μm in area) in plane of contact were quantified using qDF in real time and compared between fixed cells that had rolled to arrest between the 3 conditions (n = 3 separate experiments; **P < .01 compared with vehicle).

Figure 5.

Figure 5.

E-selectin engagement of sLe x on L-selectin requires cross-linking to induce phospho-Lck. (A) E-selectin or P-selectin cross-linked by the addition of polyclonal secondary antibody were immunoprecipitated in the absence and presence of glycomimetics. Western blot was applied to detect Lck along with its tyrosine phosphorylated state. (B) Activation of phospho-Lck was normalized to total protein to quantify the increase after E-selectin or P-selectin cross-linking in the presence of Rivipansel (6.5 μM) or GSnP-6 (12 μM). No significant increase was quantified for cross-linked P-selectin. Data presented as mean ± SEM for cross-linked E-selectin were significant compared with E-selectin alone (n = 3 separate experiments; **P < .01; ***P < .001).

Figure 6.

Figure 6.

E-selectin cross-linking activates release of MRP8/14 and TLR4 signaled extension of β 2 -integrin with upshift to a high-affinity state. (A) Diagram of outside-in mechanosignaling of β2-integrin activation via engagement of E-selectin and subsequent clustering of L-selectin on human PMNs. E-selectin binding to L-selectin induces extracellular release of MRP8/14 that then binds TLR4, which activates upshift from low-affinity to an extended intermediate-affinity state of β2-integrin. Further clustering of E-selectin bound L-selectin activates high-affinity β2-integrin. (B) MRP8/14 release by human PMNs treated with E-selectin and cross-linked in the absence or presence of Rivipansel (100 μM) or activation with IL-8 (10 nM). Release into supernatant measured by ELISA as mean ± SEM (n = 3 separate experiments; **P < .01; ***P < .001 significance, as indicated). (C) Affinity state of β2-integrin detected on PMNs in suspension by flow cytometry. β2-integrin receptors bound by KIM127 (extended intermediate affinity) versus mAb24 (high affinity) is shown for PMNs treated with E-selectin with or without cross-linker in the absence or presence of Rivipansel (100 μM), TLR4 blocking antibody, or IL-8 (10 nM; n = 3 separate experiments; #analysis of extended, *high-affinity with P < .01 compared with cross-linked E-selectin). (D) MRP8/14 (0.8 ng/mL) activation and extension of β2-integrin analyzed by flow cytometric detection of KIM127 and mAb24 receptor number in the absence and presence of TLR4 blocking antibody. Data shown as mean ± SEM (n = 3 separate experiments; #P < .05 compared with unstimulated control or TLR4 blocked samples).

Comment in

References

    1. Ley K. The role of selectins in inflammation and disease. Trends Mol Med. 2003;9(6):263-268. -PubMed
    1. Lawrence MB, Springer TA. Leukocytes roll on a selectin at physiologic flow rates: distinction from and prerequisite for adhesion through integrins. Cell. 1991;65(5):859-873. -PubMed
    1. Foxall C, Watson SR, Dowbenko D, et al. The three members of the selectin receptor family recognize a common carbohydrate epitope, the sialyl Lewis(x) oligosaccharide. J Cell Biol. 1992;117(4):895-902. -PMC -PubMed
    1. Nimrichter L, Burdick MM, Aoki K, et al. E-selectin receptors on human leukocytes. Blood. 2008;112(9):3744-3752. -PMC -PubMed
    1. Smolen JE, Petersen TK, Koch C, et al. L-selectin signaling of neutrophil adhesion and degranulation involves p38 mitogen-activated protein kinase. J Biol Chem. 2000;275(21):15876-15884. -PubMed

Publication types

MeSH terms

Substances

Grants and funding

LinkOut - more resources