The Leukocyte Adhesion Cascade: Integrin Signalling and Assays

Every leukocyte that reaches an inflamed tissue has solved the same mechanical problem. It must stop, in under a second, against a shear flow that would otherwise sweep it past the site of infection. The solution is not one adhesion event but a strictly ordered cascade — capture, rolling, chemokine-triggered arrest, spreading, crawling and diapedesis — in which each step licenses the next. The logic is unusually well defined, the human loss-of-function phenotypes are unambiguous, and almost every node is druggable.

Priming the endothelium and catching the cell

The cascade begins with the endothelium, not the leukocyte. TNF-α and IL-1β from resident macrophages drive nuclear translocation of p65 (RelA), which transcribes E-selectin, ICAM-1 and VCAM-1. The timing is the part usually glossed over: P-selectin is already made and stored in Weibel–Palade bodies, so it reaches the surface within minutes of histamine or thrombin, whereas the NF-κB-dependent proteins take two to six hours. An inflamed vessel therefore has two adhesive phases, recruiting different cells.

Capture is a mechanical trick. PSGL-1, decorated with sialyl-Lewisₓ and sulphated tyrosines, binds P- and E-selectin through catch bonds, whose lifetime increases with applied force — so rolling is stable at physiological shear and fails in a flow chamber run too gently. L-selectin adds secondary tethering onto cells already rolling and CD44 supplies a further E-selectin ligand. Ligated PSGL-1 also signals through Src kinases and Syk to push LFA-1 into an extended-closed intermediate, giving slow rolling and time to sample chemokines.

The arrest signal: chemokines and Gαi

Arrest is triggered by chemokines immobilised on endothelial glycosaminoglycans and heparan sulphate, not by soluble ones, and the receptor pairing is lineage-specific. CXCL8 and CXCL1 acting on CXCR2 dominate neutrophil arrest, CCL2 and CCL5 on CCR2 recruit monocytes, and CXCL12 via CXCR4 and interferon-induced ligands via CXCR3 direct lymphocytes. Presentation is the whole point: a surface-bound chemokine delivers a sharp, spatially confined pulse, whereas the same molecule in solution desensitises the receptor and abolishes arrest.

The receptor couples to Gαi — the step pertussis toxin abolishes, still the cleanest test that a phenotype is chemokine-driven. Gβγ activates PI3Kγ to generate PIP₃ and PLCβ2 to release DAG and Ca²⁺, which switch on CalDAG-GEFI and load Rap1 with GTP; PIP₃ recruits DOCK2 and Btk to the leading edge. Rap1-GTP, with its effector RIAM, delivers talin-1 to the membrane. Human CalDAG-GEFI deficiency phenocopies integrin deficiency despite normal integrin levels — the clearest evidence that adhesion is set by conformation, not abundance.

Inside-out signalling: talin-1, kindlin-3 and the β tails

Integrin activation is a physical event. The talin-1 head binds the membrane-proximal NPxY motif of the β2 tail (CD18) or the β1 tail (CD29) and breaks the salt bridge clamping the α and β transmembrane helices together, letting the ectodomain unbend into the extended-open state. Kindlin-3 binds the membrane-distal NxxY motif and is obligatory without being sufficient alone. PIP5K1γ supplies the PI(4,5)P₂ anchoring talin, while SHARPIN holds β2 inactive through the α tail and filamin A competes for the same site.

The human genetics separates these steps cleanly. Loss of CD18 gives LAD-I: no surface β2 integrin, delayed cord separation, no pus. Loss of kindlin-3 gives LAD-III, where selectins and integrin expression are normal but the receptor cannot be switched on, plus bleeding because platelets use the same co-activator. Activated receptors then pair predictably: LFA-1 with ICAM-1 and ICAM-2, Mac-1 with ICAM-1 plus iC3b and fibrinogen, p150,95 overlapping, VLA-4 with VCAM-1, and α4β7 with MAdCAM-1 — the reason α4β7 blockade is gut-selective.

Outside-in signalling, spreading and crawling

Once ligand is bound under force, signalling reverses direction. Clustered integrins recruit Src-family kinases, which phosphorylate ITAM-bearing adaptors and activate Syk. Syk phosphorylates SLP-76 and ADAP, scaffolding Vav1, the exchange factor for Rac1. Active Rac1 relieves autoinhibition of WASP, which drives Arp2/3 to nucleate branched F-actin and build the lamellipodium. This is the module that fails in Wiskott–Aldrich syndrome, which is why those patients have adhesion and migration defects, not a purely cytoskeletal one.

Spreading alone would immobilise the cell; crawling needs the rear to let go. RhoA acts through ROCK1 on myosin-IIA to contract the uropod and strip adhesions, FAK and Pyk2 set adhesion turnover and ezrin tethers cortical actin to the membrane. Adhesion strength versus migration speed is therefore a bell curve: cells that adhere too well crawl as badly as cells that adhere too little. SHP-1 limits how far outside-in signalling propagates, while Mac-1-dependent crawling carries the arrested cell across the flow to a permissive junction.

Diapedesis: the junction and the basement membrane

Most leukocytes cross paracellularly, through a defined sequence rather than a generic gap. PECAM-1 (CD31) engages homophilically and mobilises the lateral border recycling compartment, a subjunctional membrane reservoir that supplies the extra surface area needed to push a nucleus through. JAM-A and ESAM act at the same stage; CD99 acts later, and blocking it arrests cells part-way through the junction with the leading edge already abluminal — how the ordering was established.

The junction must also loosen: VE-cadherin is phosphorylated and displaced laterally, while claudin-5 governs the tight-junction component and, in the brain, most of the barrier. A minority instead take the transcellular route through the endothelial body, leaving VE-cadherin intact and going unnoticed if permeability is the only readout. Beyond it sit the basement membrane and pericyte sheath, breached at low-expression regions and remodelled by MMP-9 and MMP-2 — the proteases also responsible for the collateral damage of heavy neutrophil infiltration.

Key targets and matching reagents

Studying leukocyte recruitment in vivo

Adhesion is one of the few pathways where an intravenous blocking antibody gives an immediate, quantifiable readout: rolling velocity, adherent cells per field, or extravasated counts in a lavage. Experiments fall into three groups.

1. Blocking the integrin arm

Anti-CD11a (I21/7) and anti-CD18 (M18/2) define the β2-dependent component of a phenotype; anti-CD11b separates crawling and complement-receptor function from initial arrest, and anti-CD11c addresses p150,95. On the β1 side, anti-CD49d (PS/2) blocks VLA-4–VCAM-1, anti-CD29 (TS2/16) targets the shared β1 subunit, and anti-integrin β7 (FIB21) isolates the gut-selective α4β7 route. Running the β2 and α4 blocks side by side is more informative than either alone, because tissues differ in which arm dominates.

2. Removing capture and the chemokine cue

Anti-CD162 (4RA10) and anti-CD62L remove capture and secondary tethering; pairing them with anti-ICAM-1 (BE29G1) tests whether the defect sits on the leukocyte or the endothelial side. Anti-CXCR4 (12G5) and anti-CXCR3 (CXCR3-173) block lymphocyte-directing receptors, anti-CD44 (IM7) covers the extra E-selectin ligand, anti-CD31 (390) holds cells at the junction, and anti-TNF and anti-IL-1β act upstream of endothelial activation. All are low-endotoxin and functional grade — contaminating LPS activates endothelium directly and will manufacture the phenotype you are trying to block.

3. Quantifying the readouts

Blocking becomes quantitative when paired with protein measurement. Track the endothelial output with E-selectin, P-selectin, VCAM-1, ICAM-2, MAdCAM-1, VE-cadherin, claudin-5 and JAM-A kits; the chemokine cue with CXCL1, CXCL8, CXCL12, CCL2, CCL5, CCR2 and CXCR2; and the signalling module with Rap1A, talin-1, kindlin-3, integrin β2, RhoA, Vav1, phospho-SHP-1 and p65. Tissue damage tracks with MMP-9 and MMP-2.

Featured products for this pathway

Frequently asked questions

For research use only. Not for use in diagnostic or therapeutic procedures.