The Src Kinase p56 lck Up-regulates VLA-4 Integrin Affinity

In circulating lymphocytes, the VLA-4 integrin preexists in multiple affinity states that mediate spontaneous tethering, rolling, and arrest on its endothelial ligand, vascular cell adhesion molecule-1 (VCAM-1). The regulation and function of VLA-4 affinity in lymphocytes has never been elucidated. We show here that p56 lck , the major Src kinase in T cells, is a key regulator of high affinity VLA-4. This high affinity is essential for the rapid development of firm adhesion of resting T cells to VCAM-1 and to their extracellular matrix ligand, fibronectin. Lck-regulated VLA-4 function does not require intact TCR nor several key components of the TCR signaling pathway, including ZAP-70 and SLP-76. Furthermore, stimulation of p56 lck by the phosphatase inhibitor, pervanadate, triggers firm VLA-4-dependent adhesion to VCAM-1. Although Lck is not required for chemokine receptor signaling to mitogen-activated protein kinase, the presence of Lck-regulated high affinity VLA-4 also facilitates firm adhesion triggered by the chemokine, SDF-1, at short-lived contacts. Surprisingly, bond formation rates, ability to tether cells to VLA-4 ligand, and VLA-4 tether bond stability under shear flow are not affected by VLA-4 affinity or Lck activity. Thus, the ability of high affinity VLA-4 to arrest cells on VCAM-1 under flow arises from instantaneous post-ligand strengthening rather than from increased kinetic stability of individual VLA-4 bonds. These results suggest that p56 lck maintains high affinity VLA-4 on circulating lymphocytes, which determines their ability to strengthen VLA-4 adhesion and rapidly respond to proadhesive chemokine signals at endothelial sites.

Lymphocytes circulating throughout the blood and lymphatic tissues respond to inflammatory cytokines and antigenic signals (1). These cytokine and antigenic stimuli regulate both the expression and function of vascular receptors on effector and memory lymphocytes, which determine their preferential migration to extra lymphoid sites of inflammation (2). To emigrate from the circulation into inflamed tissues, lymphocytes must develop rapid firm adhesion to endothelial ligands, dis-played on the vascular endothelium within the target tissue. This firm adhesion is mediated primarily by the binding of lymphocyte integrins such as VLA-4 (␣ 4 ␤ 1 ), LFA-1, and ␣ 4 ␤ 7 to their respective endothelial ligands expressed or induced on vessel walls of target sites (3,4). To induce cell arrest on their endothelial ligands, lymphocyte integrins must rapidly stabilize their short-lived contacts with endothelial ligands without modulation of integrin surface expression. Stabilization of lymphocyte contacts could be critical for their subsequent response to endothelium-displayed chemoattractants (chemokines) (5)(6)(7), which further up-regulate integrin avidity as well as lymphocyte spreading on the endothelium, critical steps in lymphocyte extravasation to the surrounding tissue.
The ␣ 4 integrins are rather unique among other leukocyte integrins in that they maintain basal heterogeneous affinity states, which enable binding to their endothelial ligands spontaneously in the absence of cellular activation (3, 8 -13). Previously we and others have demonstrated that low affinity VLA-4 subsets mediate weak tethering and rolling interactions of VLA-4-expressing lymphocytes on vascular cell adhesion molecule (VCAM-1) 1 under shear flow (3,9,14). We further demonstrated that a small subset of high affinity VLA-4 expressed by a portion of circulating peripheral blood T lymphocytes (PBL) or resting T leukemia Jurkat cells supports tethers followed by firm arrest on high density VCAM-1, even in the absence of stimulatory chemokines (14). The mode of regulation of this high affinity subset on effector T cells by intracellular signaling pathways has not been elucidated to date.
p56 lck (Lck) is a key Src tyrosine kinase of the TCR signaling machinery of thymocytes and peripheral T lymphocytes (15)(16)(17). Lck can physically associate with CD4, CD8, CD2, and CD28 coreceptors (18 -20) as well as with several receptors to major lymphocyte cytokines such as IL-2 and IL-7 (21,22) and can be membrane-anchored within sphingolipid-rich microdomains, RAFTS (23,24). Ligation of these coreceptors and costimulatory molecules like L-selectin activate Lck even in the absence of antigenic stimuli or TCR ligation (20,25,26). Although cross-linking of TCR as well as of the aforementioned costimulatory receptors can activate integrin-dependent adhesive processes (27)(28)(29)(30), the majority of these processes require prolonged static contact with surface-ligands where the contribution of integrin clustering, cytoskeletal remodeling, and cell spreading to the generation of adhesive strength is large (31). Indeed, stimulation of integrin function by TCR-signaling elements and associated costimulatory molecules is not accompanied by induction of high affinity states in the major lymphocyte integrins VLA-4, VLA-5, and LFA-1 (31)(32)(33).
To test the involvement of the TCR machinery in modulating VLA-4 adhesion in T cells, we compared the affinity states and adhesive properties of VLA-4 on Lck-deficient and -reconstituted Jurkat cells. We demonstrate here that a high affinity VLA-4 subset continuously regulated by Lck preexists on resting Jurkat cells as well as on freshly isolated blood T cells. This subset is crucial for these lymphocytes to firmly adhere to endothelial VCAM-1 and the extracellular matrix ligand fibronectin (FN), both in the absence and presence of chemokinetriggering signals. We show that this high affinity subset is continuously regulated by Lck in resting cells, independent of their TCR activation state and of key components of the TCR signaling machinery. Lck-deficient cells are devoid of high affinity VLA-4 subsets, display reduced adhesion strengthening to VCAM-1 and FN, and exhibit impaired integrin responsiveness to T cell chemokines compared with their Lck-reconstituted analogue cells. The study also sheds new light on how integrin affinity to ligand translates into tether adhesion strengthening events under dynamic context of shear flow. Our work suggests that Lck activity in T cells is a checkpoint in the generation of high affinity VLA-4, which may determine the migratory patterns of T cells to specific sites of inflammation expressing high levels of endothelial and extracellular matrix VLA-4 ligands.

Immunofluorescence Flow Cytometry
Cells were washed once with PBS, resuspended in FACS wash buffer (PBS, 5% bovine serum albumin, and 0.05% sodium azide) and incubated with primary antibodies (10 g/ml) for 30 min at 4°C. The samples were then washed and incubated with secondary antibodies for 30 min at 4°C. Cells were washed and analyzed immediately on a FACScan flow cytometer (Becton Dickinson, Erembodegem, Belgium).
For surface staining of ␤ 1 integrin activation epitopes and for monitoring the induction of ligand-induced binding site (LIBS) epitopes by soluble peptide ligand, intact or treated cells were incubated for 5 min at 24°C in H/H binding medium (Hanks' buffered saline solution containing 2 mg/ml bovine serum albumin and 10 mM Hepes, pH 7.4 supplemented with 1 mM CaCl 2 and 1 mM MgCl 2 ) alone or in the presence of LDV-or DVL-containing octapeptides. The LIBS-specific anti-␤ 1 mAbs, 15/7, 9EG7, or HUTS, were then added at 10 g/ml to the cell suspension for an additional 5 min at 24°C, and cells were then incubated on ice for an additional 30-min period. Unbound mAbs were washed out at 4°C by multiple washing followed by staining with PE-conjugated goat-anti mouse Ig. Induction of the 15/7 or 9EG7 epitope was determined and expressed as mean fluorescence intensity units, with the preimmune mIgG control staining subtracted. To compare ligand-induced 15/7 staining among differently treated PBL, 15/7 staining in the presence of 10 g/ml control DVL peptide was subtracted from the 15/7 staining in the presence of equimolar LDV peptide.

Binding of [ 3 H]BIO-1211 to ␣ 4 ␤ 1 -expressing Jurkat Cells
[ 3 H]BIO-1211, a radiolabeled LDV derivative was prepared as described (44). For equilibrium binding measurements, Jurkat cells were incubated for 40 min with increasing concentrations of [ 3 H]BIO-1211 in Tris-buffered saline binding medium (50 mM Tris HCl, 150 mM NaCl, 0.1% bovine serum albumin, 2 mM glucose, 10 mM HEPES pH 7.4) containing either Ca 2ϩ and Mg 2ϩ (1 mM each) or Mn 2ϩ (2 mM). The cells were pelleted by centrifugation, and bound ligand was quantitated by scintillation counting as described (35). Background binding determined in the absence of divalent cations was subtracted from total binding counts. For k on measurements, cells were incubated with 0.5-5 nM [ 3 H]BIO-1211, and at specific time points, cell samples were taken, diluted with 500-fold excess of unlabeled BIO-1211 ligand, and immediately pelleted and quantitated by scintillation counting. The k on was determined by multiplying the observed binding rate constant by the BIO-1211 concentration. For k off measurements, cells were incubated with 5 nM [ 3 H]BIO-1211 for 40 min, diluted with a 500-fold excess of unlabeled BIO-1211 to quench further binding of [ 3 H]BIO-1211, and further incubated for increasing periods up to 120 min. Cell samples were pelleted at different time points, and radiolabeled ligand binding was determined as above. Data were expressed as a percentage of the maximum specific radioligand binding and were fit to an exponential curve by nonlinear regression, yielding the first order k off as described (35,44).

Soluble Bivalent (VCAM-1) 2 -Ig Binding Assay
Cells (0.5 ϫ 10 6 ) were incubated with (VCAM-1) 2 -Ig (first two domain segments of human VCAM-1 fused to a single human Fc domain) for 30 min at room temperature in H/H binding medium. Cells were then washed twice by suspension in H/H binding medium and incubated with PE-donkey anti-human IgG (Jackson) for 20 min at 4°C (in binding medium) before washing with PBS, and fluorescence was analyzed by FACS. The results were expressed as mean fluorescence intensity units.

Preparation of Adhesive Substrates
Affinity-purified seven-domain human VCAM-1 (45) or human FN (Life Technologies) was dissolved in PBS buffered with 20 mM bicarbonate, pH 8.5, and incubated on a polystyrene plate (60 ϫ 15-mm Petri dish; Becton Dickinson, Lincoln Park, NJ) for 2 h at 37°C or overnight at 4°C. The plate was then washed three times and blocked with human serum albumin (20 mg/ml PBS) for 2 h at 4°C. For SDF-1 coimmobilization on the adhesive substrate, ligands were coated in the presence of normal or heat-denatured chemokine (R&D Systems, Minneapolis, MN) (2 g/ml) and a carrier protein (human serum albumin, 2 g/ml) and washed and quenched as above.

Laminar Flow Adhesion Assays
Analysis of Cell Tethering, Rolling, and Arrest-A polystyrene plate on which purified ligand had been adsorbed was assembled in a parallel plate laminar flow chamber (260-m gap) mounted on the stage of an inverted phase-contrast microscope (Diaphot 30, Nikon, Tokyo, Japan) as described previously (46,47). Cells were washed with cation-free H/H medium containing 5 mM EDTA and stored in cation-free H/H medium at room temperature up to 1 h. Cells were diluted with H/H binding medium and perfused through the flow chamber at the desired shear stress generated with an automated syringe pump (Harvard Apparatus, Natick, MA). All experiments were carried out at 37°C. Cell images were videotaped with a long integration LIS-700 CCD video camera (Applitech, Holon, Israel) and a Sony SLV E400 video recorder (Sony, Japan) and manually quantified by analysis directly from the monitor screen. All adhesive interactions between the cells flowing over ligand-coated substrates at low shear stresses of 0.5-1.5 dyn/cm 2 were tracked and quantified. Tethering events were defined as adhesive interactions between cells and the substrate. Four categories of cell tethers under flow were defined as follows. Tethers were defined as transient if cells attached briefly (Ͻ2 s) to the substrate in a VLA-4-dependent manner, rolling if cells displayed rolling motions Ͼ5 s with a velocity of at least 1 m/s, and slow rolling if cells rolled Ͼ5 s with a velocity of less than 1 m/s, and arrests were defined as cells that immediately arrested on the substrate, remaining adherent and stationary for at least 20 s of continuous flow. Firm arrests were defined as cells that arrested under flow and resisted detachment when subjected to abruptly increased shear force of 2.5 dyn/cm 2 for 5 s. The number of tethers for each was divided by the flux of freely flowing cells in close proximity to the substrate that moved through the same field during the analysis period.
Analysis of Transient Cell Tethers to Low Density VCAM-1-Transient tethers were determined as previously described (47) at a resolution of 0.02 s in a digital still playback mode (Panasonic AG-7355, Osaka, Japan). Background tethering to VCAM-1-coated fields blocked with the VCAM-1-blocking mAb, 4B9, was less than 5% of overall tethers. All tethering events were set to start at t ϭ 0, and the natural log of the tethers that remained bound after initiation of tethering derived from the duration of all transient tethers and was plotted against tether duration to yield a slope ϭ Ϫk off .
Analysis of Cell Resistance to Detachment after Static Contact with Ligand-Cells were allowed to settle onto the substrate for 1-2 min at stasis. Flow was then initiated and increased stepwise every 5 s. The number of cells that remained bound was expressed relative to the number of cells originally settled on the substrate. Prolonged adhesion assays were similarly performed by allowing cells to settle for 15 min on flow chamber-assembled substrates. Nonadherent cells were removed by a single 2-s wash of cells at a shear stress of 50 dyn/cm 2 .

Spontaneous VLA-4 Adhesiveness to VCAM-1 and Fibronectin and Expression of the ␤ 1 Activation Epitope Is Impaired in
Lck-deficient T Cells-The adhesive properties of VLA-4 in a mutant Jurkat line that lack functional Lck were compared with those in the same mutant reconstituted with functional Lck (Ref. 16 and Fig. 1A). Both Lck-deficient and -reconstituted T cells expressed comparable levels of the ␣ 4 and ␤ 1 subunits of the VLA-4 heterodimer, the major integrin receptor on these cells for the endothelial ligand VCAM-1 and the extracellular matrix ligand FN, along with moderate but similar levels of another ␤ 1 integrin, the VLA-5 integrin (Fig. 1A). Notably, the Lck-reconstituted cells had restored basal levels of constitutive tyrosine phosphorylation strongly suppressed in the Lck depleted line (Fig. 1A). We therefore considered that Lck-regulated tyrosine phosphorylation events might regulate integrin-mediated adhesion of lymphocytes. To compare the adhesiveness of VLA-4 toward its endothelial ligand, VCAM-1, and toward its extracellular matrix ligand, FN, in both Lck-deficient and Lck-reconstituted cells, we tested the ability of both cell types to develop shear resistant adhesion after a brief static contact with the purified ligands, analyzed in parallel plate flow chamber assays. VLA-4-dependent adhesion to VCAM-1 or FN of Lck-expressing cells was strikingly greater than that exhibited by Lck-deficient cells (Fig. 1B). Lck-deficient cells remained defective in VLA-4-dependent adhesion to VCAM-1 or FN even after 15 min of contact with ligand (data not shown). VLA-4 on Lck-deficient cells was, however, structurally intact with respect to acquisition of activated phenotype, since external activation of VLA-4 with the nonphysiological affinity-stimulating mAb TS2/16 restored firm VLA-4 adhesiveness to both VCAM-1 and FN (data not shown). The strength of ␣ 4 integrin-dependent adhesion of both Lck-expressing and Lck-deficient Jurkat cells to surface-bound ␣ 4specific mAb was, however, indistinguishable (Fig. 1B), consistent with similar clustering capability of ␣ 4 integrins on both cell types.
Interestingly, Lck-reconstituted Jurkat cells displayed higher constitutive staining of the ␤ 1 subunit activation epitope 15/7 (36,48) than Lck-deficient cells in the absence of soluble integrin ligand (Fig. 1C) or divalent cations (data not shown), suggesting a higher activation state of ␤ 1 integrins in these cells than in the Lck-deficient ones. Lck-reconstituted cells also expressed higher levels of other ␤ 1 integrin activation epitopes including 9EG7 (37), HUTS4, and HUTS21 (38) than the Lckdeficient cells, despite similar levels of total ␤ 1 integrins on both cells (data not shown). Collectively, these results suggest that constitutive Lck activity in resting T cells is essential for ␤ 1 integrin activation and strong VLA-4 adhesiveness developed during short static contacts with the physiological integrin ligands VCAM-1 and FN.
Affinity of VLA-4 to Soluble Ligands Is Compromised in Lck-deficient Cells-We considered that the dramatically reduced adhesiveness of VLA-4 on the Lck-deficient cells could be the result of an altered intrinsic affinity of the integrin to ligand. VLA-4 recognizes an LDV-containing sequence on the CS-1 domain of FN (49) that is homologous and isosteric with the tetrapeptide Gln-Ile-Asp-Ser (QIDS), the VLA-4 binding site in VCAM-1 (50). LDV octapeptides serve as useful probes of VLA-4 affinity for both VCAM-1 and FN and can block high affinity VLA-4 (14,36). Since direct binding of monovalent VCAM-1 to cell surface VLA-4 cannot be monitored because of its extremely low binding affinity in physiological medium (33), three approaches were taken to compare VLA-4 affinity states in Lck-deficient and Lck-reconstituted Jurkat cells: 1) LDV peptide saturation binding to VLA-4 on Lck ϩ and Lck Ϫ Jurkat populations, monitored by the peptide induction of LIBS on the ␤ 1 integrin subunit of VLA-4; 2) direct binding of a radiolabeled high affinity LDV peptide derivative, BIO-1211 (44) (Table I). Notably, the difference in VLA-4 affinity between the Lck-expressing and Lck-deficient cells toward BIO-1211 was much smaller than the difference in apparent VLA-4 affinity (i.e. EC 50 ) between Lck-expressing and Lck-deficient cells toward the LDV octapeptide. This reflects the 10 6 -fold higher binding affinity of BIO-1211 versus the LDV octapeptide to VLA-4 (35,44). Interestingly, the lower K d values (higher affinity) of BIO-1211 bound to VLA-4 on Lck-expressing cells were exclusively due to 2-fold slower dissociation rates of the ligand from high affinity VLA-4, with no change in the K on values between the higher and lower affinity VLA-4 states. Ligand dissociated more slowly from VLA-4 on Lck-expressing cells than from VLA-4 on Lck-deficient cells even in the presence of Mn 2ϩ , even though or Lck-reconstituted (closed squares) cells to VCAM-1 (medium density; 0.2 g/ml) or fibronectin (high density; 5 g/ml) or to the ␣ 4 integrin-specific mAb HP1/2 (coated at 0.2 g/ml on the substrate) at stasis. Strength of adhesion was analyzed by measuring the resistance to detachment by controlled shear forces of compared Jurkat populations settled for 2 min on ligand-coated substrates assembled as the lower plate of a flow chamber. Cells were then subjected to an incremented shear stress assay. The number of cells remaining bound at the end of each 5-s interval of shear increase was expressed relative to the number of cells originally settled on the substrate. Antibody blocking with the anti-␣ 4 HP1/2 mAb completely eliminated T cell adhesion to VCAM-1 (Fig. 1B, left panel) and reduced adhesion to FN by Ͼ90% (data not shown) in these short term static assays. All measurements were performed at 37°C in adhesion binding medium. The mean Ϯ range of two independent experiments is depicted. Results are representative of six and three independent experiments, respectively. C, FACS staining of the ␤ 1 activation epitope 15/7 in Lck-deficient (thin line) and-reconstituted (thick line) Jurkat cells.
this cation prolonged VLA-4 bond lifetime by 100-fold relative to physiological cations (Table I). The increased affinity of ligand recognition by VLA-4 in Lck-expressing cells is thus due to an inherently slower dissociation rate of ligand from VLA-4 in these cells. Consistent with a markedly stronger binding of soluble ligands to VLA-4 on Lck-reconstituted cells, Lck-recon-stituted cells exhibited saturated binding of bivalent VCAM-1, with an EC 50 of 1.4 nM, whereas the Lck-deficient cells bound the same ligand with an EC 50 of Ͼ10 nM (Fig. 2B and data not  shown).
Lck-triggered High Affinity VLA-4 Is Essential for Firm Cell Adhesion to VCAM-1-To determine how the presence of high affinity VLA-4 affects rapid adhesion developed by T cells on VCAM-1, we utilized LDV peptides as selective blockers of high affinity VLA-4 states (14). Lck-expressing or -deficient Jurkat cells were allowed to adhere briefly to VCAM-1 in the presence of control DVL or increasing concentrations of LDV octapeptide. The strong adhesion of Lck-expressing cells (40% Lck ϩ versus 8% Lck Ϫ ) was highly sensitive to the presence of LDV peptide, indicating the involvement of high affinity VLA-4 molecules. By contrast, the low level of VLA-4 adhesiveness to VCAM-1 developed by Lck-deficient cells was completely insensitive to inhibition by LDV peptide, indicating that these cells lack high affinity VLA-4 (Fig. 3). Consistent with these results, a soluble VCAM-Ig fusion protein at a concentration saturating the high affinity subset (10 nM) completely blocked all VLA-4dependent adhesion of Lck-expressing to VCAM-1-coated substrates but did not affect the weak adhesion developed by Lck-deficient cells on identical substrates (data not shown). This demonstrates that high VLA-4 affinity maintained by Lck activity plays an indispensable role in early VLA-4-mediated adhesion-strengthening events of Jurkat cells adhered to VCAM-1 at stasis.
We next compared VLA-4 tethering to VCAM-1 in shear flow in Lck-deficient and -reconstituted Jurkat cells. The ability of VLA-4 to tether cells to the integrin ligand did not depend on the presence of Lck (Figs. 4A). However, Lck-reconstituted Jurkat cells arrested on high density VCAM-1 within fractions of seconds and developed firm shear-resistant adhesion much more readily than Lck-deficient cells under shear flow (Fig.  4A). The high affinity VLA-4 subsets on Lck-expressing cells were the exclusive mediators of firm arrest to VCAM-1, as evident by the complete susceptibility of firm cellular arrests but not of weak arrests, rolling, or transient tethering events to soluble LDV at 12 M, a concentration selective for full occu- . One experiment is representative of three. B, bivalent (VCAM-1) 2 -Ig binding to Lck-deficient and -reconstituted Jurkat cells. Cells were incubated with increasing amounts of (VCAM-1) 2 -Ig, washed, incubated with PE-anti human IgG, and analyzed by FACScan. Background binding, determined in the presence of 5 mM EDTA, was less than 10% of the total binding and was subtracted from that binding. One experiment is representative of three.  pancy of high affinity VLA-4 (data not shown). Taken together these results suggest that high affinity VLA-4 present exclusively in Lck-expressing cells is critical for their ability to rapidly develop firm spontaneous adhesion to VCAM-1 both under shear flow and during short static contact. Immunolocalization of VLA-4 by electron microscopy indicated similar distribution of the integrin on surface microvilli in both Lckdeficient and -reconstituted cells (data not shown), suggesting that VLA-4 topography is Lck-independent. To further show that the adhesive defect in Lck-deficient cells is affinity and not integrin topography-dependent, we substituted anti-␣ 4 integrin-specific antibodies for VLA-4 ligands as the adhesive substrate. Lck-deficient cells showed no defect in their ability to tether and arrest on a surface-immobilized anti-␣ 4 integrin mAb HP1/2 under shear flow, consistent with the similar adhesion to these mAbs after short static contact ( Fig. 1B and  data not shown).
The ability of high affinity VLA-4 subsets to rapidly arrest Lck-expressing cells on high density VCAM-1 under flow could reflect stronger adhesive bonds. To test this possibility, we compared transient VLA-4-mediated tethering of Lck-expressing or Lck-deficient cells to low density VCAM-1 under shear flow. In this assay, the contribution of integrin rearrangement during formation of transient tethers to low density VCAM-1 is minimized since the majority of tethers lasts Ͻ0.5 s (Fig. 4B), too rapid for diffusion-limited integrin rearrangements (51). In agreement with the similar tethering frequency mediated by VLA-4 in Lck-deficient and -reconstituted cells on physiological high densities of VCAM-1 (Fig. 4A) and with the similar k on measured for the binding of the LDV derivative to VLA-4 on both cell types (Table I), the formation rate of the quantal adhesive tethers mediated by high or low affinity VLA-4 to low density VCAM-1 were found to be similar (Fig. 4B). Transient tethers dissociated from low VCAM-1 with a single exponential corresponding to a single first order dissociation rate constant (k off ) (Fig. 4B). This dissociation rate constant varied with the shear stress but did not increase upon further dilution of VCAM-1 (not shown), suggesting it represented quantal adhesive units. Surprisingly, and in contrast to the 2-fold lower k off measured for high affinity VLA-4 bonds than for low affinity VLA-4 bonds (Table I) and to the markedly higher binding of soluble VCAM-1 to high affinity VLA-4 on Lck-reconstituted cells, no significant difference was noted in the k off of VLA-4 tethers mediated by the high or low affinity VLA-4 on Lckreconstituted or Lck-deficient cells, respectively. This suggests that shear flow conditions abolish k off differences of tether bonds that are evident in equilibrium binding measurements. Nevertheless, although not contributing to stronger adhesive bonds formed between VLA-4 and very low density VCAM-1 under shear flow, high affinity VLA-4 recognition of VCAM-1 translated robustly within subseconds to adhesion strengthening events on high density ligand both under flow (Figs. 4A) and in stasis (Fig. 3).
Regulation of VLA-4 Activity by Lck Occurs Independently of the TCR Machinery-It is well established that TCR stimulation can modulate integrin adhesiveness (27,28,52). Lck plays a prominent and upstream role in TCR stimulation; Lck phosphorylates signaling motifs, known as immunoreceptor tyrosine-based activation motifs, which recruit the Syk family tyrosine kinase, ZAP-70, leading to phosphorylation of a large number of signaling proteins, including the adaptor protein, SLP-76. Therefore we asked whether the constitutive Lck-dependent VLA-4 adhesiveness implicates key components of the TCR machinery. To test this hypothesis, we analyzed VLA-4 function in mutant Jurkat T cell lines lacking expression of the TCR ␤ chain, ZAP-70, or SLP-76. Each mutant was compared with a paired reconstituted clone stably transfected with the wild-type cDNA and expressing comparable levels of ␣ 4 integrin subunit of VLA-4 (Fig. 5, A-C). Analysis of the ability of each pair to develop shear resistant adhesion to VCAM-1 showed that the absence of the TCR, ZAP-70, or SLP-76 did not impair VLA-4 adhesive function nor did it effect the activation state of the integrin, as detected by the ␤ 1 epitope-specific mAb 15/7 (Fig. 5, A-C). This is in direct contrast to the highly impaired level of mAb 15/7 activation epitope and markedly reduced VLA-4-mediated adhesion in Lck-deficient cells (Fig. 1,  B and C). Together, these results rule out a role for the TCR and other key TCR components in the constitutive Lck-dependent up-regulation of VLA-4 function.
Lck-mediated Activation of VLA-4 Depends on the Kinase Activity of Lck but Not ZAP-70 -Given our finding that Lckmediated regulation of VLA-4 occurs independently of its central role in TCR signaling, we next asked whether the kinase activity of Lck is required for this function. We used the tyrosine phosphatase inhibitor, pervanadate (PV) to augment Lck The motions of all tethered cells were analyzed on high density VCAM-1, coated at 0.5 g/ml, and grouped into four categories of tethers as described under "Experimental Procedures." The fraction of each category is presented in each stacked bar. One representative experiment of four is shown. B, transient tethering of Lck-deficient and -reconstituted cells to low density VCAM-1 (coated at 0.02 g/ml) at low shear stress of 0.5 dyn/cm 2 . Kinetics of formation and dissociation of VLA-4-mediated tethers, formed by equal number of Lck-deficient or -reconstituted Jurkat cells perfused over the substrate for 1 min. The natural log of tethers with duration lower or equal to the indicated time as described under "Experimental Procedures." is shown. Residual number of tethers longer than 1 s were not fitted to the first order dissociation curve. In A and B, more than 95% of all tethers were blocked by T cell pretreatment with the anti-␣ 4 integrin mAb, HP1/2 (not shown). One experiment is representative of three. kinase activity (53)(54)(55)(56)(57). Treatment of cells with pervanadate induced tyrosine phosphorylation of multiple cellular proteins in an Lck-dependent manner, since it was blocked by the Src kinase inhibitor PP2 and was not observed in Lck-deficient cells (Fig. 6A). Pervanadate treatment of Lck-expressing Jurkat cells resulted in increased VLA-4 adhesion to VCAM-1 during rapid static contact but did not affect VLA-4 adhesion of Lck-deficient cells to VCAM-1 (Fig. 6B). In contrast, direct DAG-dependent PKC activation with phorbol 12-myristate 13acetate could bypass Lck in triggering VLA-4 adhesiveness (Fig. 6B). To test the specific role of Lck kinase activity in regulating VLA-4 function, we tested the contribution of ZAP-70 protein-tyrosine kinase in pervanadate-induced upregulation of VLA-4-mediated adhesion. Notably, ZAP-70-deficient and -reconstituted cells displayed similar pervanadateinduced increases in VLA-4-mediated adhesion (Fig. 6C), suggesting that ZAP-70 is not required, nor does it modulate pervanadate-induced Lck-mediated up-regulation of VLA-4 adhesion. In addition, PV stimulation of VLA-4 activity in Lckexpressing T cells was inhibited by the selective DAG-dependent PKC inhibitor GF 109203X (data not shown), suggesting that Lck activation of VLA-4 is mediated by DAG-dependent PKC activity. High affinity VLA-4 expressed by resting Lckexpressing Jurkat cells could also be suppressed upon cell treatment with GF 109203X, further indicating that both spon-taneously maintained and PV-stimulated high affinity VLA-4 requires intact PKC activity.
Jurkat cell integrin regulation by T cell activation signals resembles that of effector VLA-4 hi PBL (14). Consistent with a major role of Lck activity in VLA-4 activation in Jurkat cells, short pervanadate stimulation of freshly isolated human PBL triggered LDV induction of VLA-4-associated LIBS, which was completely sensitive to PP2 inhibition (Fig. 7A). Pervanadate also enhanced immediate VLA-4-dependent lymphocyte arrests and strengthened lymphocyte rolling adhesions on  ). B, extent of cell adherence and resistance to detachment by shear flow developed by differently treated Lck-deficient or -reconstituted Jurkat cells settled on medium density VCAM-1 (coated at 0.1 g/ml). Spontaneous or agonist-stimulated VLA-4-dependent adhesion is depicted. Cells were left intact or pretreated for 2 min with either phorbol 12-myristate 13-acetate (PMA, 100 ng/ml) or PV (100 M) before being settled for 2 min on the VCAM-1 substrate. Cells were then subjected to an incremented shear stress assay as described in Fig. 1. No adhesion to VCAM-1 of either type of Jurkat cells was observed if cells were pretreated with the anti-␣ 4 subunit mAb HP1/2 (not shown). The mean Ϯ range of two independent experiments is depicted. Results are representative of four independent experiments. C, extent of cell adherence developed by resting or agonist-treated ZAP-70-deficient or -reconstituted Jurkat cells settled on medium density VCAM-1 (coated at 0.1 g/ml). All treatments and assays are as described in B. The mean Ϯ range of two independent experiments is depicted. Results are representative of three independent experiments.
VCAM-1 under physiological shear flow (data not shown). Short exposure of T cells to IL-2 stimulates Lck activity independent of antigenic stimulation (21,58,59). Consistent with this finding, short pretreatment of PBL with IL-2 enhanced firm VLA-4 adhesion to VCAM-1 under physiological flow relative to that of untreated PBL (Fig. 7B). Relatively low concentrations of LDV peptide, which failed to interfere with any adhesive interactions mediated by VLA-4 on untreated cells, could completely inhibit the IL-2-augmented firm VLA-4-mediated adhesion to VCAM-1 (Fig. 7B). Transient VLA-4-mediated adhesions, which were not affected by the IL-2 stimulation, were also not susceptible to inhibition by the LDV peptide. Whereas VLA-4 on Lck-deficient Jurkat showed no response to IL-2 stimulation, Lck-reconstituted cells did develop firm adhesion to VCAM-1 under flow upon short exposure to the cytokine (data not shown). Thus, IL-2-triggered rapid VLA-4 adhesion strengthening to VCAM-1 is Lck-dependent and is mediated by LDV-susceptible high affinity VLA-4 subsets on T cells.
The Lck-dependent High Affinity VLA-4 Subset Facilitates Chemokine-stimulated T Cell Adhesion-Chemokines can rapidly trigger leukocyte adhesion to vascular integrin ligands both under shear flow and during short static contact (6,60). Chemokine stimulation of integrin adhesion requires intact G i protein but is independent of DAG-dependent PKC (61). We wished to examine whether the Lck-dependent high affinity VLA-4 on T cells plays any regulatory role in the integrin response to chemokine stimulation by SDF-1, a prototype T cell chemokine that strongly triggers VLA-4 adhesiveness under physiological flow (62). Lck-reconstituted and Lck-deficient Jurkat cells express comparably high levels of the SDF-1 receptor CXCR4 (Fig. 8A), which was functional in both cellular systems, triggering equivalent G-protein signaling activity in an Lck-independent manner (Fig. 8B). Consistent with this, Lck activity was not required for optimal SDF-1-induced signaling activity of CXCR4 in normal T cells, since the Src kinase inhibitor PP2 did not affect SDF-1-triggered mitogen-activated protein kinase activation in Lck-reconstituted cells (Fig. 8B). Nevertheless, whereas SDF-1 induced robust VLA-4 adhesion of Lck-expressing cells interacting with either VCAM-1 or FN, it failed to stimulate VLA-4-dependent adhesion in Lck-deficient cells (Fig. 9A). To test whether the defect in chemokine responsiveness of the Lck-deficient cells was directly linked to impaired VLA-4 affinity state, we artificially activated VLA-4 affinity on Lck-deficient cells by the integrin-activating mAb TS2/16. Indeed, integrin affinity-stimulated Lck-deficient Jurkat cells were rendered responsive to SDF-1 stimulation of VLA-4 adhesion (Fig. 9B). Lck-deficient cells were also responsive to SDF-1 stimulation of VLA-4 adhesion to high density VCAM-1 (data not shown), suggesting that the VLA-4 avidity and contact strength rather than intrinsic affinity determine the ability of VLA-4 to respond to chemokine. These experi- ments suggest for the first time that proadhesive effects of chemokines on integrin function are facilitated by integrin avidity to ligand and that preexistent high affinity subsets can augment chemokine triggering of integrin adhesion. DISCUSSION Integrin-mediated adhesion strengthening can result from changes in intrinsic affinity to ligand as well as alterations in lateral integrin clustering and post-ligand cell spreading, which generate high avidity to multivalent ligand through low affinity integrin subsets (31,32,63,64). Leukocyte adhesion to vascular endothelium under shear flow involves the establishment of rapid integrin-dependent adhesion within subseconds to seconds of leukocyte contact with the endothelium, suggesting that high affinity recognition of ligand by integrin molecules is important rather than integrin recruitment at the adhesive zone or cell spreading (31,32). In this study, we examined how integrin affinity contributes to cell adhesion in different contexts of recognition of immobilized ligands, i.e. ligand density, contact time, shear flow, and proadhesive cytokines. Our study assigns a major role for high affinity integrin recognition of ligand in regulating rapid development of firm integrin-dependent adhesion at short-lived adhesive zones under physiological flow and during short static contacts.
A subset of VLA-4, found by us to be constitutively regulated on T cells by the Src kinase p56 lck , exhibits high affinity to both fibronectin-derived VLA-4 binding peptides and soluble VCAM-1. This high affinity VLA-4 was found crucial for development of firm VLA-4-mediated adhesions by T cells interacting with VCAM-1 and fibronectin at short-lived contacts and under shear flow. Adhesive interactions mediated by low affinity VLA-4 subsets were found to support only weak adhesions to VCAM-1, which readily dissociated under high shear forces (Figs. 1B, 3, and 4A). Dissecting the relationship between ligand binding properties and adhesiveness of distinct states of VLA-4, we found that both high affinity and low affinity VLA-4 subsets (expressed by Lck-expressing and -deficient cells, respectively) bind a soluble monovalent fibronectin-derived VLA-4 peptide with similar k on , which corresponded to similar rates of tether formation supported by both VLA-4 subsets on immobilized VCAM-1 (Fig. 4). These results are consistent with our previous findings that the formation rate of VLA-4 tethers to VCAM-1 does not depend on the VLA-4 affinity state in Jurkat and PBL (14). High affinity VLA-4 recognition of VCAM-1 in Lck-expressing cells, but not in Lck-deficient cells could, however, translate within subseconds of contact to rapid adhesion strengthening on high density VCAM-1 under flow (Fig. 4A). Unexpectedly, high affinity VLA-4 dissociated from low density VCAM-1 with similar rates to low affinity VLA-4 ( Fig. 4B), suggesting that intrinsic tether bond lifetimes mediated by either low or high affinity VLA-4 are similar. Similar findings were observed with another Jurkat cell line, an activation mutant defective in PKC signaling and high VLA-4 affinity that failed to translate VLA-4 tethers to high density VCAM-1 into adhesion strengthening events (14) but exhibited identical bond formation rates and similar kinetic stability of tether bonds in the presence of shear flow. 2 This is a first demonstration that equilibrium properties of integrin bonds do not correlate to the kinetics of these bonds in the presence of tensile shear forces. Nevertheless, despite its negligible contribution to stability of individual VLA-4 tether bonds to low 2 C. Chen and R. Alon, unpublished results.  FN. A, resistance to detachment by shear flow developed by Lck-deficient or -reconstituted Jurkat cells settled for 2 min on low density VCAM-1 (0.05 g/ml) or fibronectin (2 g/ml) alone or in the presence of coimmobilized SDF-1 (2 g/ml). Note that at low density integrin ligand in the absence of chemokine, both types of Jurkat cells failed to develop detectable adhesion. No adhesion of either cell type to the VCAM-1 or FN coimmobilized with chemokine was observed in the presence of the ␣ 4 integrin-blocking mAb, HP1/2, or the integrin inhibitor EDTA, respectively (data not shown). The mean Ϯ range of data collected from two representative fields of view are depicted. One experiment is representative of three. B, resistance to detachment by shear flow developed by Lck-deficient Jurkat cells settled on VCAM-1 alone or coimmobilized with SDF-1 after artificial activation of VLA-4 affinity with the ␤ 1 integrin-activating mAb TS2/16. Cells were preincubated in binding medium alone or with 0.5 g/ml mAb for 5 min and allowed to settle for 2 min on low density VCAM-1 (0.05 g/ml) before being subject to detachment. TS2/16-stimulated adhesion to VCAM-1 in the absence of presence of chemokine was integrin-specific as it was completely blocked in the presence of EDTA. One experiment is representative of three. density VCAM-1 under flow, high affinity VLA-4 has the capacity to instantaneously translate its tethers to firm adhesion on high density VCAM-1 (Fig. 4A). When this high affinity subset is functionally suppressed (by Lck deficiency) or blocked (by soluble LDV peptide), the ability of VLA-4 tethers to develop rapid adhesion strengthening to VCAM-1 under flow or upon static contact with ligand is completely lost. The unique ability of high affinity VLA-4 to nucleate firm arrest under flow could not be attributed to increased ability of high affinity VLA-4 molecules to rearrange and cluster with ligand at subsecond contact zones; both high and low affinity, VLA-4 on Lck-reconstituted or Lck-deficient cells, respectively, shared a comparable ability to tether to and cluster on immobilized VLA-4 binding mAbs (Fig. 1B and data not shown). We conclude that high affinity occupancy of the integrin is obligatory to generate instantaneous adhesion strengthening on multivalent ligand at subsecond contact zones under shear flow and firm shear-resistant adhesion at short static contacts.
Lymphocyte adherence to endothelium can be rapidly upregulated upon encounter with endothelium-displayed chemokines (6,62,65). We have recently found that immobilized chemokine presented to adherent cells in juxtaposition to VLA-4 or LFA-1 ligands can rapidly augment integrin avidity to these ligands both at stasis and under shear flow (62,65). We now report that high affinity VLA-4 expressed on Lck-expressing cells plays a crucial regulatory role in this process in that these cells preferentially develop firm adhesion in response to surface-bound chemokines. Chemokine signaling is, however, Lck-independent, as SDF-1 induced rapid extracellular signalregulated kinase 1/2 phosphorylation in Lck-deficient cells as in Lck-reconstituted cells (Fig. 8B). Chemokine triggering of VLA-4 avidity results in G i -protein-dependent integrin clustering at the adhesive contact zone and is not associated with alteration of integrin affinity to ligand (62). Rather, the cells use preexistent affinity states of VLA-4 to elicit various types of adhesive interactions with its ligands in response to adjacent chemokine signals. Thus, the ability of high affinity VLA-4 to nucleate high avidity interactions with ligand not only facilitates firm chemokine-independent VLA-4 adhesion to ligand but also augments the responsiveness of VLA-4 on interacting cells to proadhesive chemokines presented to them at adhesive contact sites. Indeed, above a threshold of integrin ligand density, even low affinity VLA-4 on Lck-deficient cells could efficiently respond to coimmobilized chemokine (data not shown). This suggests that increased contact avidity to high density ligand can compensate for the loss of high affinity VLA-4 in Lck-deficient cells and rescue the poor responsiveness of VLA-4 on these cells to immobilized chemokine. Further investigation is under way to test if this novel role of high affinity integrin to function as a nucleation element of both spontaneous and chemokine-triggered VLA-4 avidity to ligand could apply to other vascular integrins such as LFA-1 and Mac-1. Recent findings indicate that VLA-4 integrin ligation can increase LFA-1 avidity to ICAM-1 at static contact (66). High affinity VLA-4 is thought to be essential for this VLA-4 cross-talk with LFA-1, suggestive of yet another function of high affinity VLA-4 subsets in controlling vascular adhesion of circulating lymphocytes to inflamed endothelium expressing both VCAM-1 and ICAM-1.
Lck maintains basal levels of constitutive activity in resting T cells (59), regulating a steady state level of tyrosine phosphorylation on a large number of lymphocyte proteins. Consistent with tight regulation of VLA-4 function by p56 lck , short inhibition of Jurkat cell or blood T lymphocyte phosphatases by pervanadate directly activates p56 lck (53-57) concomitantly with activation of VLA-4 adhesion. Recently, the two major T cell Src kinases Fyn and Lck were shown to have partially overlapping but distinct functions in Jurkat cells (67). The complete lack of effect of PV on VLA-4 activity in Lck-deficient Jurkat cells demonstrates that Lck is an indispensible regulator of high affinity VLA-4 in these T cells. Our results suggest a minor if any compensatory role for other Src kinases, such as Fyn, at the level of which they are expressed and active in Jurkat cells (16,67). Of course, our results do not rule out the possibility that other Src kinases may play some role in regulating VLA-4 affinity in other cell types. Additionally, the full effect of PV in triggering VLA-4 activity in ZAP-70-deficient cells, impaired in their activation of PLC␥1 in response to PV (42), suggests that Lck activation of VLA-4 function does not require and is not modulated by ZAP-70 or PLC␥1 activities. Although we link Lck-regulated tyrosine phosphorylation to VLA-4 affinity, tyrosine phosphorylation of integrin tail cytoplasmic domains per se is unlikely to be implicated in Lckregulated VLA-4 affinity. Integrin tyrosine phosphorylation does not result in altered binding to soluble ligand (68, 69) but, rather, has been implicated in physical association with the cytoskeleton (70). Furthermore, activation of VLA-4 function through p56 lck required intact activities of DAG-dependent PKC (data not shown), implicating serine/threonine phosphorylation events in Lck-regulated VLA-4 function. It has been shown that Lck localizes within discrete lipid microdomains, RAFTS, (24) which are enriched in heterotrimeric G proteins and TCR signaling elements (71). As VLA-4 colocalizes to specific RAFTS (72), these domains may serve as platforms for Lck to sequester and activate affinity modulators of VLA-4. Candidate Lck substrates for induction of high VLA-4 affinity subsets are members of the Ras family, shown to modulate ␤ 1 integrin affinity to ligand in fibroblasts (73,74) and T cells (75). Rap1, a Ras-family GTPase, has been recently shown to regulate integrin adhesiveness in T cells, but it does not appear to control constitutive VLA-4 avidity to ligand (76).
The dramatic differences in VLA-4-dependent adhesion to VCAM-1 caused by small changes in binding affinity on subset of VLA-4 molecules shown by us suggests that VLA-4 affinity in circulating T cells plays a crucial Lck-regulated function in versatile rapid adhesion strengthening processes mediated by this integrin. p56 lck activity is subjected to complex intracellular regulation by both antigen-dependent TCR ligation as well as by antigen-independent exposure to inflammatory cytokines, shown to induce integrin activation epitopes that correlate with high affinity states on effector lymphocytes (77). Indeed, short T cell exposure to IL-2 is sufficient to enhance rapid T cell adhesion to VCAM-1 under shear flow mediated by high affinity VLA-4 subsets (Fig. 7). Prolonged activation of p56 lck on effector T cells by IL-2 (22) is therefore likely to up-regulate both VLA-4 affinity and rapid adhesion strengthening on ligand at dynamic contexts. TCR activation of VLA-4 adhesion by antigen (27)(28)(29) and costimulatory elements physically associated with Lck (19,20,30,78) likely involve Lck activation as well. Lck may thus be a key translator of both cytokine and antigenic information transmitted to effector T cells at lymphoid tissues before they exit these sites and return to the circulation. Lck-dependent acquisition of firm T cell adhesiveness to VCAM-1, a key homing receptor of effector cells at sites of inflammation, may serve as a checkpoint in the ability of these lymphocytes to emigrate from blood vessels at inflamed extra-lymphoid target organs. to thank G. Cinamon for assistance in PBL isolation and Dr. S. Schwarzbaum for editorial assistance. Special thanks to Drs. A. Peled and R. Seger for helpful discussions.