INTRODUCTION

Intercellular adhesion molecule 1 (ICAM-1,¹ CD54) is a heavily glycosylated, single-chain 80-114-kDa protein that is composed of five extracellular Ig-like domains, a transmembrane spanning region, and a cytoplasmic tail (1, 2). It is a ligand for LFA-1 (CD11a/CD18), Mac-1 (CD11b/CD18), CD43, rhinovirus, and fibrinogen, and it participates in cellular interactions by binding to its several ligands via different domains. ICAM-1 is expressed constitutively at low levels on lymphocytes, vascular endothelium, and a variety of other cell types (3, 4). In vivo, high levels are expressed on tissues involved in inflammatory responses (5, 6). In vitro, ICAM-1 expression is rapidly up-regulated by inflammatory cytokines such as interferons, IL-1, and tumor necrosis factor . Its expression on lymphocytes is up-regulated by antigen recognition (3, 7).

The use of blocking monoclonal antibodies against LFA-1 and ICAM-1 has established that the LFA-1/ICAM-1 interaction plays a major role in a variety of adhesion-dependent leukocyte functions and immune responses. We have shown that anti-ICAM-1 blocks T-dependent B cell activation (8) and cytotoxic T-lymphocyte target recognition and activation (9, 10), and inflammatory responses (4, 11) are also inhibited. It was originally thought that LFA-1 and ICAM-1 acted solely as adhesion molecules, strengthening otherwise weak interactions between cells. However, there is increasing evidence to suggest that ICAM-1 plays a signaling role. We showed that co-cross-linking of ICAM-1 (CD54) and MHC II induced expression of a functional IL-2 receptor on murine B cells (8). In addition, ICAM-1 cross-linking has been shown to cause an oxidative burst in neutrophils (12) and to modulate anti-IgM induced changes in intracellular Ca²⁺ in a Burkitt's lymphoma cell line (13). Furthermore, an anti-ICAM-1 mAb modulated the release of interferon-, tumor necrosis factor , and IL-1 in T lymphocytes and monocytes (14). These results support the idea that ICAM-1 is involved in signal transduction processes.

Signal transduction by cell surface receptors is regulated by changes in the activity of specific kinases and/or phosphatases. The structure of ICAM-1 does not predict intrinsic tyrosine kinase activity, therefore if ICAM-1 is to transduce signals, specific cytoplasmic tyrosine kinases must associate with the receptor. Among candidate receptor-associated protein-tyrosine kinases are members of the Src-related family of cytoplasmic protein-tyrosine kinases, which includes several critical signaling molecules that have been shown to mediate signaling function both in lymphocyte development and in antigen responses (15, 16). Also, tyrosine phosphorylation of the 34-kDa cdc2 protein kinase has been shown to be transiently induced in response to ICAM-1 cross-linking in T cells (17).

The activation of receptor-associated tyrosine kinases can lead to downstream activation of the mitogen-activated protein (MAP) kinase pathway (18, 19), including extracellular regulated kinase-1 (ERK) and ERK-2. These mediate a number of functional cell changes, such as cell growth, differentiation, and gene induction (20), and are known to be activated in leukocytes in response to stimuli.

We describe here the activation of the Src-related protein-tyrosine kinase p56/p59^lyn and the involvement of Raf-1 and MAP kinase in ICAM-1 signaling within the B lymphoma line A20.

EXPERIMENTAL PROCEDURES

mAbs used were P7/7.1 (rat IgG_2b anti-murine MHC II) (21), MAR 18.5 (mouse IgG_2a anti-rat ) (22), 145/2C11 (hamster IgG anti-murine CD3) (23), 187.1 (rat IgG_2b anti-murine ) (24), PY72.10.5 (mouse IgG₁ anti-phosphotyrosine) (25), YN1/1.7.4.1 (rat IgG_2b anti-murine CD54/ICAM-1) (26), and M1/42 (rat IgG_2a anti-murine MHC I) (27). Monoclonal antibodies were affinity-purified from culture supernatants using Protein G-Sepharose (Pharmacia Biotech Inc., Montreal, Quebec, Canada). Polyclonal anti-Lyn, anti-Fyn, anti-Lck, and anti-MAPK antisera were obtained from Upstate Biotechnology (Lake Placid, NY) and anti Raf-1 antiserum from Santa-Cruz Biotechnologies. PMA was obtained from Sigma.

The B cell line A20 was derived from a BALB/c lymphoma (IgG⁺, IgM⁺, IgA) (28). Cells were grown in RPMI 1640 medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 5 × 10⁵ M 2-mercaptoethanol, penicillin (100 units/ml), and streptomycin (100 µg/ml). LG1 is a BALB/c (Ia^d) CD4⁺ VB6⁺ T cell line that is specific for ovalbumin. When required, T cells were incubated at 1 × 10⁶ cells/ml overnight in anti-CD3-coated microwells (10 µg/ml in phosphate-buffered saline). The cells were fixed at a concentration of 1 × 10⁷ cells/ml by incubation in 0.8% paraformaldehyde (Sigma) for 5 min, followed by incubation in 0.2 M lysine (Sigma) for 1 min.

B cells were incubated at 5 × 10⁷/ml at 37 °C in culture medium for 10 min. mAbs, paraformaldehyde-fixed T cells, or PMA (Sigma) were added where appropriate and incubated for various amounts of time at 37 °C. Cells were then cooled on ice, washed three times with Dulbecco's phosphate-buffered saline + 1 mM NaVO₄, and lysed by incubation in lysis buffer (10 mM Tris-HCl, pH 7.5, 37 mM NaCl, 1% Nonidet P-40, 1 mM NaVO₄, 2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and 50 mM NaF) for 15 min on ice. Cell debris was removed by centrifugation (10 min at 13,000 rpm). The protein content of each lysate was quantified using a bicinchoninic acid assay (Pierce, Aurora, Ontario).

Lysates were boiled for 5 min with an equal volume of 2 × sample buffer containing 5% 2-mercaptoethanol before loading. Equal amounts of protein were loaded in each lane. Following electrophoresis, separated proteins were then transferred to nitrocellulose membranes (Bio-Rad, Mississauga, Ontario) at 100 V for 1 h. Membranes were blocked for 1 h in blocking buffer (5% bovine serum albumin fraction V (Boehringer Mannheim, Laval, Quebec)) in wash buffer (10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 2.5 mM EDTA, pH 8.0, 0.1% Tween). Membranes were incubated with PY72.10.5 (1 µg/ml) in blocking buffer for 2 h followed by ¹²⁵I-goat anti-mouse Ig (ICN, Mississauga, Ontario) for 1 h. Radioactivity was revealed by autoradiography and quantified by PhosphorImager analysis (Molecular Dynamics, Inc.). Alternatively, membranes were probed with various primary antibodies and detected using the ECL system with horseradish peroxidase-conjugated secondary antibodies (Amersham Corp., Oakville, Ontario) according to the manufacturer's protocol.

Lysates were precleared twice by incubation at 4 °C with Pansorbin (Calbiochem) for 30 min. Lysates were then incubated with 5 µg of anti-Lyn, anti-Fyn, and anti-Lck antisera followed by a 30-min incubation with Pansorbin. Pansorbin beads with bound protein were then washed three times in lysis buffer, and proteins were extracted by incubation in loading buffer for 15 min at room temperature. Samples were cleared by centrifugation (30 s at 13,000 rpm), and supernatants were loaded onto SDS-PAGE gels or used for kinase assays.

After immunoprecipitation, Pansorbin beads with bound protein were washed three times in lysis buffer with 1 mM NaVO₃, once with lysis buffer with 1 M NaCl, once with lysis buffer without inhibitors, and finally once with kinase buffer (20 mM HEPES, pH 7.5, 100 mM NaCl, 5 mM MnCl₂, and 5 mM MgCl₂). The samples were then resuspended in 25 µl of kinase bufffer containing 1 µM cold ATP and 1.5 µl of [³²P]ATP and incubated for 5 min at room temperature. Pansorbin beads with bound protein were then washed three times in lysis buffer, and proteins were extracted by incubation in loading buffer for 15 min at room temperature. Samples were cleared by centrifugation (30 s at 13,000 rpm), and supernatants were loaded onto SDS-PAGE gels. The gels were fixed overnight in a 10% acetic acid, 30% methanol solution and were dried, visualized by autoradiography, and quantified by PhosphorImager analysis.

Staining was performed with 20-min incubations of 5 × 10⁵ cells/tube at 4 °C with biotinylated mAb at concentrations ranging between 2 and 20 µg/ml, with phycoerythrin-coupled streptavidin (PharMingen, Mississauga, ON). Fluorescence was analyzed using a FACScan (Becton Dickinson, Mississauga, Ontario). Cell debris was excluded on analysis by side scatter gating.

RESULTS

Contact with activated CD4⁺ T cells is a crucial step in B cell activation. To investigate the intracellular signaling events mediated by contact with activated T cells, we examined changes in protein tyrosine phosphorylation. A20 cells were coincubated for 1 h with resting or anti-CD3-stimulated paraformaldehyde-fixed T cells (LG1). Anti-phosphotyrosine immunoblotting of whole cell lysates showed an increase in protein tyrosine phosphorylation of A20 incubated with activated T cells that was greater than the sum of the activated T and B cells alone (Fig. 1A). This was not the case when A20 were incubated with resting T cells. The fact that the T cells were fixed excludes a role for soluble cytokines. In addition, supernatant from activated LG1 did not have this effect (data not shown). Thus the tyrosine hyperphosphorylation seen was mediated by ligation of B cell surface receptors by coreceptors on activated T cells. In previous studies we had shown that ICAM-1/LFA-1 interactions make a critical contribution to contact signaling for B cells (29) and therefore asked whether ICAM-1 signaling could contribute to the observed tyrosine phosphorylation. Increased protein tyrosine phosphorylation was detected when anti-ICAM-1 antibodies were cross-linked on A20 cells; this was detectable as early as 1 min (Fig. 1B). In particular, hyperphosphorylation of proteins in the molecular mass range of 50-60 kDa was observed. Similar results were obtained using another lymphoma line, TA3 (data not shown). Tyrosine hyperphosphorylation was not seen when either of the isotype-matched anti-MHC I (M1/42) or anti-Thy-1 (30H12) antibodies were substituted for anti-ICAM-1.

0

Cross-linking ICAM-1 on splenic B cells up-regulated MHC II expression (8). Analogous treatment of A20 cells induces similar MHC II expression (Fig. 2), showing that ICAM-1 has a signaling function in A20 cells analogous to that already described in non-transformed B cells.

The proteins that were hyperphosphorylated following ICAM-1 signaling included some with molecular weights corresponding to those of the Src family of kinases. Upon activation, these kinases are autophosphorylated on tyrosine residues. We therefore immunoprecipitated several of these kinases (p53/p56,^lyn p59^fyn, and p56^lck) from lysates of A20 cells that were either unstimulated or stimulated through treatment with cross-linked anti-ICAM-1. The immunoprecipitates were separated by SDS-PAGE and immunoblotted with anti-phosphotyrosine antibodies. Tyrosine phosphorylation of a pair of bands specifically precipitated by anti-p53/p56^lyn was increased approximately 2-fold following ICAM-1 cross-linking, according to PhosphorImager analysis (Fig. 3). Greater than 1.2-fold hyperphosphorylation of the other kinases was not seen.

An in vitro kinase assay verified that p53/p56^lyn kinase was activated. This occurred within 1 min following ICAM-1 cross-linking and peaked at approximately 10 min (Fig. 4A). The quantitation of this experiment is shown in Fig. 4B.

0

The finding that ICAM-1 ligation resulted in tyrosine phosphorylation of several cellular proteins, including Lyn kinase, suggested that the Ras-Raf-1-MAPK/ERK kinase-MAPK pathway might also be activated, as changes in tyrosine phosphorylation induced by cell surface receptors have been implicated in this pathway of MAP kinase activation. The activation of both Raf-1 and MAPK is accompanied by a characteristic shift in their mobility on SDS-polyacrylamide gels (30, 31). Immunoblotting of cell lysates with anti-Raf-1 detected a protein with a molecular mass of 72 kDa, and a shift in electrophoretic mobility was detected within 1 min of ICAM-1 cross-linking (Fig. 5A). Immunoblotting with an anti-MAPK antibody detected two proteins of molecular masses 42 and 44 kDa, corresponding to ERK-2 and ERK-1, respectively (Fig. 5B). There was a shift in the electrophoretic mobility of ERK-1 within 1 min following ICAM-1 stimulation, similar to that seen in cells stimulated with the protein kinase C activator PMA. This was not observed when an isotype-matched control antibody, anti-MHC I (M1/42), was cross-linked on A20 cells indicating that the effect was not due to Fc-receptor engagement and was specific to ICAM-1. We find that M1/42 does not induce tyrosine hyperphosphorylation in A20 cells, so this lack of effect on MAPK was expected.

0

DISCUSSION

Adhesion molecules of several families can transduce signals that influence the regulation of cell growth and differentiation. The ₂ integrins have been the most extensively studied (reviewed in Ref. 32). We have now investigated the mechanism of signal transduction through ICAM-1 in a murine B cell lymphoma line, A20. The interaction between ICAM-1 and LFA-1 is known to be important in B and T cell activation (4, 9-11), and while there is evidence for a signaling function for ICAM-1 (8, 12, 13), relatively little is currently known about the mechanism. We have now shown the activation, in B lymphoma cells, of the Src family kinase p53/p56^lyn, as well as Raf-1 and MAP kinases.

The fact that ligation of ICAM-1 resulted in MHC class II up-regulation confirmed that this cell surface molecule transduces biological signals in A20 cells. ICAM-1 cross-linking also induces rapid tyrosine hyperphosphorylation of a number of proteins, in both A20 cells and in another B lymphoma line, TA3. Similar protein tyrosine phosphorylation was induced in A20 by contact with fixed, activated T cells, and given that ICAM-1 is critical for this mode of B cell activation (29), it is likely that ICAM-1 signaling contributes. We determined that one of the proteins that becomes hyperphosphorylated on tyrosine is p53/p56^lyn kinase. Lyn kinase has been detected in various hematopoietic cells, including B cells (33), neutrophils (34), and eosinophils (35). Two isoforms of the Lyn-encoded protein have been identified, p53 and p56, arising from differential splicing. Src-related kinases are autophosphorylated during the process of activation. An in vitro kinase assay confirmed that p53/p56^lyn was activated as early as 1 min after ICAM-1 cross-linking. In contrast, p59^lyn and p56^lck were not phosphorylated upon tyrosine following ICAM-1 cross-linking.

Tyrosine kinase activation resulting from such diverse stimuli as cytokines, growth factors, and T cell receptor ligation leads to the activation of MAP kinases (18, 19). Members of the MAP kinase family are protein/serine/threonine kinases that require dual phosphorylation on threonine and tyrosine residues for full activation (36, 37). A unique type of dual threonine/tyrosine kinase known as MAPK/ERK kinase is primarily responsible for phosphorylation and activation of ERK. These dual specificity kinases are themselves activated by phosphorylation (38, 39). The kinase Raf-1 has been shown to activate MAPK/ERK kinase (40). Several mechanisms have been implicated in the regulation of Raf-1, including its tyrosine phosphorylation by activated growth factor receptors (41), serine/threonine phosphorylation by protein kinase C (42), and physical interaction with the GTP-binding Ras protooncogene family (43-45). We found that ICAM-1 cross-linking resulted in a shift in the electrophoretic mobility of Raf-1 and ERK-1, indicating an activation of these kinases. We propose that ICAM-1 ligation allows association with and/or activation of Lyn resulting in tyrosine phosphorylation of cellular proteins. Whether this functional association is indirect via other as yet undefined intermediates or occurs via generation of SH2 ligand phosphotyrosines on ICAM-1 itself remains to be determined. While the cytoplasmic domain of ICAM-1 contains two tyrosine residues, it does not contain the common tyrosine-containing motif ((D/E)X₇(D/E)X₂YX₃LX₇X₂(L/I)) that has been shown to mediate Src family kinase binding to other receptors (46). Nonetheless, Lyn has been demonstrated to bind to the common  chain of the granulocyte-macrophage colony-stimulating factor, IL-3, and IL-5 receptors, which also lack this motif (35).

Chirathaworn et al. (17) showed tyrosine hyperphosphorylation in primary and transformed human T cells stimulated with anti-ICAM-1 antibody. They identified as a substrate the cell cycle regulator cdc2 kinase and showed a corresponding transient inhibition of cdc2 kinase activity, implicating ICAM-1 in growth arrest. Although we have shown phosphorylation of substrates more commonly associated with cell cyle progression, preliminary experiments also show that ICAM-1 cross-linking inhibits growth of A20 cells. The role of activated Raf-1 and MAPK in such growth inhibition remains to be determined.

The downstream effects of ICAM-1-mediated activation of Lyn, Raf-1, and MAP kinases are as yet unknown. In other systems, MAP kinases activate many cytosolic proteins such as phospholipase C, phopholipase A2, and nuclear transactivating factors such as nuclear factor IL-6, c-Fos, c-Myc, p62^tcf, and others (37), thus mediating a wide range of activation and differentiable events. Raf-1 has been shown to activate nuclear factor-B by directly phosphorylating its inhibitor, IB (47). ICAM-1 ligation induces up-regulation of MHC class II, up-regulation of cytokine receptors (8), modulation of B cell receptor signaling (13), and induction of an oxidative burst in neutrophils (12). The biochemical signaling pathway that we have now identified may be involved in these responses. A future goal is to define the role of this and other signaling pathways in evoking specific cellular responses.