Integrin-mediated tyrosine phosphorylation and cytokine message induction in monocytic cells. A possible signaling role for the Syk tyrosine kinase.

Activation of cytoplasmic tyrosine kinases is an important aspect of signal transduction mediated by integrins. In the human monocytic cell line THP-1, either integrin-dependent cell adhesion to fibronectin or ligation of β1 integrins with antibodies causes a rapid and intense tyrosine phosphorylation of two sets of proteins of about 65-75 and 120-125 kDa. In addition, integrin ligation leads to nuclear translocation of the p50 and p65 subunits of the NF-κB transcription factor, to activation of a reporter gene driven by a promoter containing NF-κB sites, and to increased levels of mRNAs for immediate-early genes, including the cytokine interleukin (IL)-1β. The tyrosine kinase inhibitors genistein and herbimycin A block both integrin-mediated tyrosine phosphorylation and increases in IL-1β message levels, indicating a causal relationship between the two events. The components tyrosine phosphorylated subsequent to cell adhesion include paxillin, pp125, and the SH2 domain containing tyrosine kinase Syk. In contrast, integrin ligation with antibodies induces tyrosine phosphorylation of Syk but not of FAK or paxillin. In adhering cells, pre-treatment with cytochalasin D suppresses tyrosine phosphorylation of FAK and paxillin but not of Syk, while IL-1β message induction is unaffected. These observations indicate that the Syk tyrosine kinase may be an important component of an integrin signaling pathway in monocytic cells, leading to activation of NF-κB and to increased levels of cytokine messages.


Integrin-mediated Tyrosine Phosphorylation and Cytokine Message Induction in Monocytic Cells
A POSSIBLE SIGNALING ROLE FOR THE Syk TYROSINE KINASE* (Received for publication, January 31, 1995, and in revised form, March 16, 1995) Tsung H. Linz, Carlos Rosalesr, Krishna  Activation of cytoplasmic tyrosine kinases is an important aspect of signal transduction mediated by integrins. In the human monocytic cell line THP-l, either integrin-dependent cell adhesion to fibronectin or ligation of /31 integrins with antibodies causes a rapid and intense tyrosine phosphorylation of two sets of proteins of about 65-75 and 120-125 kDa. In addition, integrin ligation leads to nuclear translocation of the p50 and p65 subunits of the NF-KlJ transcription factor, to activation of a reporter gene driven by a promoter containing NF-KlJ sites, and to increased levels of mRNAs for immediate-early genes, including the cytokine interleukin (IL)-I/3. The tyrosine kinase inhibitors genistein and herbimycin A block both integrin-mediated tyrosine phosphorylation and increases in IL-l/3 message levels, indicating a causal relationship between the two events. The components tyrosine phosphorylated subsequent to cell adhesion include paxillin, pp125 F AK , and the SH2 domain containing tyrosine kinase Syk. In contrast, integrln ligation with antibodies induces tyrosine phosphorylation of Syk but not of FAI{ or paxillin. In adhering cells, pre-treatment with cytochalasin D suppresses tyrosine phosphorylation of F AI{ and paxillin but not of Syk, while IL-l/3 message induction is unaffected. These observations indicate that the Syk tyrosine kinase may be an important component of an integrin signaling pathway in monocytic cells, leading to activation of NF-KlJ and to increased levels of cytokine messages.
Members of the integrin family of cell surface receptors are involved in many key biological processes, including cell to cell and cell to extracellular matrix adhesion, cell motility, hemostasis, lymphocyte trafficking, and inflammatory phenomena (1)(2)(3), Integrins are comprised of noncovalently linked a/{3 heterodimers (4). Different ex and {3 subunits can associate in various combinations, which then determine the ligand-binding specificities of the intact integrin heterodimer complexes (2,4). Recently, it has become apparent that integrins function not only as adhesive proteins but can also transduce biochemical signals into the interior of the cell (5). One mode of integrin * This work was supported by grants from the National Institutes of Health (to S. H. and R. L. J.) and by a grant from the American Cancer Society(to R. L. J.). The costs ofpublication ofthis article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. signal transduction involves the activation of cytoplasmic tyrosine kinases. In fibroblasts, platelets, endothelial cells, and cultured tumor cells, integrin-induced tyrosine phosphorylation involves a novel cytoplasmic tyrosine kinase termed pp125 focal adhesion kinase (ppI25 F AK ) 1 (6)(7)(8). However, other aspects of integrin signaling have been observed, including calcium transients (9), changes in cytoplasmic pH (10), modulation of ion channels (11), activation of protein kinase C (12), as well as numerous other effects (13). The relationship between these later events and integrin-mediated tyrosine phosphorylation is currently unclear.
There have been observations in a variety of cell types showing that interactions with the extracellular matrix (ECM) can modulate gene expression (14)(15)(16)(17). During inflammation, blood monocytes respond to chemotactic factors and subsequently extravasate into inflamed tissues. Monocytes migrate through the subendothelial basement membrane and underlying interstitial structures rich in extracellular matrix proteins and also interact with vascular endothelial cells and connective tissue cells. Partly as a consequence of cell-cell and cell-ECM interactions, monocytes are induced to secrete cytokines and to undergo maturation to macrophages. Integrins are prominent among the cell surface receptors that mediate many of the adhesive functions of monocytes (1,2). In human peripheral blood monocytes, cell adherence to ECM components or ligation of (31 integrins with antibodies results in the rapid induction of multiple inflammatory mediator genes including several cytokines (18)(19)(20). The 5' -regulatory regions of many of the genes induced by integrin ligation contain binding motifs for the NF-KB transcription factor, suggesting a role for this factor in the gene induction process (5). In parallel to gene induction, a rapid and profound increase in protein tyrosine phosphorylation is observed, with the predominant phosphorylated component(s) having a molecular mass of about 76 kDa (21). Both of these responses are blocked by tyrosine kinase inhibitors, suggesting an important role for protein tyrosine phosphorylation in integrin signaling pathways, leading to inflammatory mediator gene induction. Aside from enhanced tyrosine phosphorylation, little is known of integrin-mediated signal transduction in monocytic cells. A significant reason for this is the difficulty involved in performing biochemical or molecular studies on peripheral blood monocytes as well as the donor to donor variability observed with these cells.
In this report, we describe a cell culture model using the human monocytic leukemia cell line, THP-l, which mimics important aspects of monocyte responses to ECM proteins. THP-l cell adhesion to ECM proteins or ligation of f31 integrins with antibody induces protein tyrosine phosphorylation, increases inflammatory mediator gene message levels, and activates the NF-KB transcription factor. These responses are blocked by tyrosine kinase inhibitors such as herbimycin A and genistein. Among the proteins tyrosine phosphorylated in THP-l cells in response to cell adhesion to ECM components are the pp125 F AK (6,22), the focal contact protein paxillin (23), and the nonreceptor tyrosine kinase Syk (24,25). Ligation of f31 integrins with intact antibodies or with F(ab')2 fragments results in the tyrosine phosphorylation of Syk but not of FAK or paxillin. The tyrosine phosphorylation of Syk is accompanied by an increase in its kinase activity. These results indicate that, like FAK, the Syk kinase is an integrin-responsive nonreceptor tyrosine kinase. They also suggest that activation of Syk is closely correlated with the induction of inflammatory mediator gene messages, while there is no such correlation for FAK. Thus, Syk may be a vital part of an integrin signaling pathway in monocytic cells, leading to transcription factor activation, and to increased levels of cytokine messages.

EXPERIMENTAL PROCEDURES
Materials-Monoclonal antibodies reactive with FAK, Raf-l, PTPID (Syp), and paxillin, as well as goat anti-mouse IgG-peroxidase and goat anti-rabbit IgG-peroxidase conjugates, were purchased from Transduction Laboratories (Lexington, KY). The anti-Sf integrin antibody P4CI0 was obtained from Life Technologies, Inc. Anti-Syk kinase polyclonal antibodies were raised in rabbits using fusion proteins as previously described (26). The mouse hybridoma TS2/16 (anti-B'l integrin subunit) was a generous gift of Dr. M. Hemler (Dana Farber Cancer Research Institute, Boston, MA). TS2/16 F(ab')2 fragment was prepared by proteolytic digestion using a kit from Pierce following the manufacturer's directions. Removal of intact antibody or Fe fragments was accomplished using a protein G affinity column. The purity of the F(ab')2 was evaluated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) prior to use in experiments. Herbimycin A, genistein, and calyculin A were purchased from Calbiochem. Protein G-Sepharose was from Pharmacia Biotech Inc. Human fibronectin, collagen type I, collagen type IV, and laminin, and the tissue culture reagents were from Life Technologies, Inc. Other reagents and chemicals were from Sigma.
Cell Culture-THP-l cells were maintained in RPMI 1640 medium containing 10% fetal calf serum (heat inactivated), 50 ILM 2-mercaptoethanol, 50 ILg/ml streptomycin, and 50 units/ml penicillin. Substratum-coated dishes were prepared by incubating 10 ILglml fibronectin, collagen, or laminin in phosphate-buffered saline (PBS) in tissue culture dishes at 4°C overnight. The dishes were blocked with 0.1 % bovine serum albumin (BSA) and washed with PBS prior to use. Adhesion Assay-Adhesion studies were performed using a modification of a previously described assay (27). Briefly, 48-well tissue culture plates were coated overnight at 4°C with 0.2 ml of PBS containing 10 ILg/ml ECM proteins. The wells were blocked with 0.1% BSA in PBS for 2 h at 4°C and washed with PBS prior to use. THP-l cells were washed, resuspended in RPMI 1640 medium, and added to substratumcoated wells (l x 10 5 cells/well) for 30 min at 37°C. The wells were washed two times with warm PBS and stained with a solution containing 0.1 % crystal violet and 10% methanol in PBS for 15 min. Following three washes with PBS, the crystal violet stain was solubilized with 1% SDS and quantitated by measuring the absorbance at 540 nm. The absorbance from 1 X 10 5 stained cells was used to calculated the number of adhesive cells per well. Each assay was performed in triplicate.
Cell Adherence and Integrin Ligation-THP-l cells were harvested from suspension culture, washed extensively with cold RPMI 1640 medium, and suspended in medium with 0.1% BSA. Cell adherence was initiated by adding the cells to 100-mm tissue culture dishes coated with ECM proteins. Cells were incubated at 37°C for the times indicated in the figure legends. For integrin ligation, THP-l cells were incubated for 45 min at 4°C in RPMI 1640 medium or medium containing intact anti-B'l IgG, or F(ab')2 fragment, washed twice with cold medium, and then incubated at 37°C in RPMI 1640 medium. Preparation of Cell Lysates-Cells were lysed in a buffer containing 50 mM Tris, pH 7.5,150 mM NaCl, 2 mM EDTA, 0.5 mM EGTA, 1 mM sodium vanadate, 0.2 !LM calyculin A, 5 mM NaF, 5 mM sodium pyrophosphate, 2 mM phenylmethylsulfonyl fluoride, 10 ILglml aprotinin, 0.5% Triton X-I00, and 0.1% sodium deoxycholate, and the lysates were cleared by centrifugation at 30,000 x g for 30 min at 4°C. Protein concentration in the lysates was determined using the bicinchonic acid assay (Pierce). Immunoblotting-Total celilysates from equivalent cell numbers or immunoprecipitated proteins were separated by SDS-PAGE (8%) under reducing conditions. The proteins were transferred eletrophoretically onto polyvinylidene fluoride membranes (Immobilon P, Millipore Corp.). The membranes were blocked with 1% BSA and 0.1% Tween 20 in PBS. The membranes were subsequently probed with primary antibody (1 ILg/ml) in PBS containing 1% BSA and 0.1% Tween 20. The antibody-antigen complexes were detected by using goat anti-mouse IgG or goat anti-rabbit IgG peroxidase conjugates, followed by use of an enhanced chemiluminescence kit (Amersham Corp.) according to the manufacturer's instruction. In some cases, the blots were stripped of bound antibodies by incubating the membranes with stripping buffer containing 62.5 mM Tris, pH 6.7, 100 mM 2-mercaptoethanol, and 2% SDS for 30 min at 50°C. The stripped blots were reprobed with other antibodies.

Immunoprecipitation and Syk Kinase Autophosphorylation Assay-
Cell lysates were precleared by incubation with protein G-Sepharose. The cleared lysates were first incubated with anti-FAK, anti-Raf-l, anti-paxillin, anti-PTPID, or anti-Syk kinase antibody for 3 h at 4°C, followed by the addition of protein G-Sepharose, and then incubated for additional 3 h at 4°C. The precipitates were washed extensively with lysis buffer. For Western analysis, the precipitates were boiled with SDS-PAGE sample buffer to dissociate the proteins. For Syk kinase autophosphorylation assay, the anti-Syk immunocomplexes were further washed twice with kinase assay buffer (50 mM HEPES, pH 7.6, 10 mM MnCI 2, 2 mM MgCI 2, and 1 mM p-nitrophenyl phosphate) and resuspended in 50 ILl ofthe same buffer containing 10 ILCi of['Y-32PJATP (3000 Cilmmol) and 2 ILM ATP. After 10 min at 30°C, the reactions were stopped by addition of 20 ILl SDS-PAGE sample buffer (4X) and boiling for 3 min.
RNA Isolation and Northern Analysis-Total cellula r RNA was isolated by th e guanidium isot hiocyanate-phenol-chloroform extraction method (34). Northern ana lysis was performed according t o a publish ed procedure (21). Bri efly, total RNA was ele ctrophore sed on 1.2% agaroseform ald ehyd e gels, tran sferred onto polyvinyl iden e fluorid e membran es (Immobilon N, Millipore), and fixed t o the membrane by UV irrad iation . Human eDNA probes for IL-l{3 (20)

THP -I Cells Adhere to ECM Component s via Integrins-We
surveyed the expression of integrins on th e THP-l cell surface by flow cytometry. THP-l cells maintained in sus pens ion culture expressed sever al inte gri n subunits including f31 , f32 , a2, a3 , a4, and as (data not shown). We also examined whether these cells could adhe re to ECM compon ent-coated substrata . As shown in Fig. lA , THP-l cells adhe red to tis su e culture plates coated with fibron ectin, laminin, collagen typ e I, or collagen type IV. THP-l cells exhibited the highest level of cell adhesion to the fibron ectin sub stratum, a moderate level of adherence to laminin, and lower levels of cell adhes ion to collagen type I or collagen type IV, whil e only a few cells adhe red to albumin-coated control wells. To test whether cell adhesi on to fibronectin was specifically med iated by integrins, THP-l cells were tr eated with RGDS peptide or with a mouse an tibody recognizing the f31 integrin subunit . Fig . 1B shows that both RGDS peptide and anti-f31 integrin antibody inhibited cell adhesion to fibronectin, whereas treatment wit h RGES peptide or normal mouse IgG had no effect . These results indicate that THP-l cells us e f31 integrins to interact with and adhere to ECM compon ents such as fibron ectin.
En gagement of THP-l Cell Integrins Increases Protein Tyrosin e Phosphorylation and IL-lf3 Message Levels-THP-l cells were plated on fibronectin-coated dishes for differ ent period s of time (7.5-60 min). Lysates from adherent cells wer e exa mined for protein phosphorylation on tyro sin e resid ues by using antiphosphotyrosine immunoblotting. As shown in Fig. 2A (top  panel), cell adhesion to fibron ectin gave ris e to a ma rked increa se in tyrosine pho spho rylation of severa l protein s, including three bro ad bands centered around 67, 71, a nd 120 kDa . Th e increas e in tyrosine pho sphorylation was readily det ect ed 7.5 min after cells were plated on to the fibronectin substratum, re ached a maximal level afte r 15 min, and then declin ed , but wa s still significa ntly elevated after 1 h . Cell adhesion to fibronectin also induced increased mes sage levels of t he inflam-was he d at 60°C th ree tim es with was hi ng bu ffer (40 mMsodi um phosphate, pH 6.8,1 mMEDTA, and 1% SDS). Th e blots were vis ua lized by exposing the membran es to films at -80°C.  m id d le panel ) shows that m aximal in duction of IL-1{3 messa ge was reach ed 1 h after cell adhesion to fibronectin began. By comp aring the time courses , it can be seen that the induction of tyro sin e pho sphorylation occurred pr ior to IL -1{3 messag e expre ssion. Th e adhesioninduced increa se in IL-1{3 message over ba sal levels was a ppro xim at ely 4 -5-fold . We h ave observed that cert ain oth er cyt okine messa ges (e.g. IL-S ) a re a lso in cr ea sed upon THP-1 adhe sion to fibron ect in (dat a not shown ).
Antibody-m ed iated ligation of integrins has been previou sly used to mimic the integrin clu stering proc ess th at occurs durin g formation of a dhe sive contacts (35). Nonadherent THP-1 cells wer e treated with the a nti-{31 integrin antibody TS2 /16 and then a nalyzed for tyro sine pho sphorylation a nd messag e levels. As shown in Fi g. 2B (top panel) , ligation of {31 integrins resulted in increa sed tyro sin e pho sphorylation (lan e 3), a lt h oug h the pattern was some wh at different from tha t induced by cell a dhesion t o fibronectin (lan e 2 ). With a nt ibody ligation , bands a t about 115 an d 67 kD a were most prominent. Ligation of {3 1 in t egrins with an tibody cou ld al so in cr ea se IL -1{3 messag e levels (Fig. 2B , middle panel, lan e 3 ). Th e observed induction of tyro sine pho sphorylation a nd message expression by a nt i-{31 a nt ibody was not du e to engageme nt of Fe r eceptors, since TS2/16 F(ab ' )2 induced tyrosin e pho sphorylation an d IL-1{3 m essage as effect ive ly a s in t act antibody (Fig. 2C ).
Integrin Ligation Leads to A ctivation of NF-KB -To det ermine wh ether integrin ligation might lead to activa tion of transcription factors a nd th e initiation of transcription, we em ployed gel mobility sh ift assays , as well as r eporter gene assays . Since the IL-1{3 promot er , as well as the promoter s for se veral ot he r monocyt e immediate-early genes, con tains NF-KB motifs (5), we decid ed to focus on act ivat ion of the NF -KB transcription complex. THP-1 cells wer e transfect ed with a pla sm id containing se ver al copies of the NF -KB motif driving a CAT r eporter gene, by a simila r plasmid with mutated NF-KB sit es , or by a lu ciferas e r eporter const ru ct driven by the c-fos promoter. As seen in Fig. 3A , THP-1 cell a dhesion to fibr onectin r esulted in a substant ial induction of CAT activity in cells transfected with the construct containing NF-KB motifs but not in cells transfect ed with the mutated NF -KB cons truct. Expression of lu ciferase driven by the c-fos promoter was not a ffecte d by cell adhesion , indicating that the results obser ved with the NF-KB-dri ven vect or were not du e to gen er ali zed in cr ea ses in transcriptional act ivity . As see n in Fig . 3B , act iva t ion of the NF-KB-driven reporter could be trigger ed by integrin ligation with intact a nt i-{31 a nt ibody or with F(ab ' )2 fragments of the a nt ibody, sugges t ing that inte gr in ligat ion, r ather than cell a dhesion, wa s su fficient for reporter act iva t ion.
Integrin ligation by ant ibodies or integrin-mediated adhesion to fibronectin a lso ca used act iva tion and nuclear transloca ti on of the p50 an d p6 5 subuni ts of th e NF-KB comp lex. As seen in Fig . 3C , using a probe containing the MH C cla ss I enhan cer NF -KB sit e, within 1 h ther e was a str ong in crea se in bands r ep r esenting p50 a n d p65 comp lexes ; an even more robu st effect wa s observed a fte r overnight st imu lation . Use of a probe containing the IL-1{3 NF-KB motif gave similar results (da t a not sh own ). Th ese observations, along with the reporter gen e assay s, sug ges t that inte grin liga tion can act iva te NF-KB , allowing this factor to cont ri bu te to the st imu lat ion of tran-  24 h after tr ansfection, cells (3 X 10 6 ) were resuspended in 1% BSA in RPMI medium and left in suspension or plated on fibronectin-coated 60-mm dishes. NF-KB-transfected cells were also stimulated with 5 ug/ml anti-131 integrin TS2/16 mAb IgG or TS2/16 Ftab ' ); fragments followed by F(ab' )2 goat anti-mouse Ab for 60 min at 4°C and then incubated at 37°C (panel B ). Cells were collected 24 h lat er and lysed to measure CAT activity or 6 h later for luciferase activity. c-fos-transfected cells were also stimulated with 20 ng/ml phorbol 12-myristate 13-acetate, and luciferase activity was measured 6 h later. In A, solid bar s represent non-adherent cells, hat ched bars represent cells adhered to Fn, . stippled bars represent cells tr eated with phorbol ester. Dat a are :!: S.E. Panel C depicts the results of gel shift assays using a probe containing NF-KB motifs from the class I MHC enhancer; the experimenta l conditions are the same as Pan els A and B . The identification of the NF-KB p50/p65 heterodimer was confirmed by supershifting the p65 band and blocking the p50 band with specific antibodies (32,33). The faster moving band was similarly identified as p50/p50 homodimer. Cells were analyzed after 1 h (la nes 1-5 ) or after overnight stimulation (lanes 6-10) under the following conditions. Lanes 1 and 6, adherence to  Herbimycin A and Genistein Inhibit Tyrosine Phosphorylation and IL-1 f3 Message Expr ession Induced by Cell Adhesion or by Ligation of f31 Integrins-To evaluate whether tyrosine phosphorylation plays a critical role in the signaling pathways leading to increased message levels , THP-l cells were treated with protein-tyrosine kinase inhibitors an d examined for effects on tyrosine phosphory lation and IL-lf3 message in duction. Herbimycin A an d genistein are selective inhibitors of tyrosine kin ases with distin ct mechanisms of action; herb imycin A blocks tyrosine kinases by attacking critical sulfhydry l groups (36), while genistein in hibits t hese enzymes by binding to t he ATP binding sites (37). Fig. 4A (top panel) demons trates that t reatme nt of THP-l cells wit h herbimycin A resul ted in an in hibition of the tyrosine phosphorylati on induce d by cell adhesion to fibronectin. Th e inhi bition was dose dependent ; 10 JLM herbimycin A strongly suppresse d the in duction of tyrosine phosphorylation, whil e 2 JLM was less effective. Expression of IL-lf3 mRNA was inhibited by herbimycin A in a manner similar to the effect on tyrosine phosphorylation (Fig. 4A, middle  pan el). Another tyrosine kinase inhibitor, genistein, also exhibited simila r inhibitory effects on tyrosine phosphorylation and IL-l f3 expression (Fig. 4B) ; t he inhibitory effects of genistein wer e noticeable at 20 JLM an d virtually complete at 100 JLM. As in the case of adherence-induced events, the res ponses induced by anti-f31 antibody could be blocked with herbimyci n A or genistein. As shown in Fig. 4C, TS2/16-induced tyrosine phosphorylation a nd IL-lf3 expression were also inhibited by her bimycin A or genistein in a dose-dependen t man ner . Thus, the res ponses induced by integrin ligati on or by adhesio n could be inhibited by the tyrosine kin ase inhibitors her bimycin A or genistein, suggesting that integrin-me diated increases in IL-lf3  Fig. 4. Syk kin ase was immunoprecipit ated from cell lysates a nd a na lyze d for it s enzy me activi ty as describ ed under "Experiment al P rocedures" (top pan el ). In parall el with the kin ase assay, the amount of Syk kin ase in th e pr ecipitated immunocompl exes was examine d by Western blot (bottom panel ). Her., 10 Ji M herbim ycin A; Gen. , 100 Ji M genis te in. message lev els re quire protein-tyrosine kinase activity.

Identification of Proteins Tyr osine Phosphorylated in Response to Cell Adhesion to a Fibronectin Substratum or to
Ligation of {31 Int egrins with Antibodies-We sought to id entify proteins th at were tyrosine phosphorylated in THP-1 cells subseque n t to engage me nt of integrins, which might be important in integrin-mediated message induction . We obtained antibodies to several proteins known to be involved in sign al transduction cascades or in cyt oskelet a l organization and examined their tyro sine pho sphorylation statu s. We examined FAK (125 kDa) as a possible component of th e 120-kDa complex and Ra f-1 (74 kDa), PTPlD (72 kD a ), pa xillin (68 kD a ), a nd Syk kin a se (72 kDa) as possib le components of the 65 kD a -75-kD a comp lexes . Sp ecific antibodies were used to immunoprecip itate these pot ential substrates , followed by immunoblotting with an ti-phosphotyr osine a nt ibody. As shown in Fig. 5A, it is clear that FAK , paxillin , and Syk kinase were tyrosine phosphorylate d u pon cell adhesion to fibronectin ; th ere wa s no evidence for tyrosine phosphorylation ofPTP1D or Raf. Only Syk kinase, but not FAK or pa xillin, wa s found to be tyrosine phosphorylated wh en cells were stimula te d by treatment with anti-{31 integrin antibody. The induction tyrosine phosphorylation of Th e pr ecipitated immunocompl exes were a na lyzed for ph osphotyrosyl-contain ing protein s by a nti-phosphoty rosine (anti· P¥) irnmunoblot ting. Th e blots were stri pped a nd reprobed wit h th e resp ecti ve antibody used for immunoprecipitation . C, cells were treated as describe d in pan el B for 1 h . Tot al cellular RNA was prob ed with IL-1(3or (3-actin probes (6 Jig total RNAIlan e). NAD, non adhe re nt; ADH, a dherent to fibronectin.
Syk kinas e wa s not du e to engageme nt of Fe r eceptors, since the F(ab ' )2 fragment of TS2/16 was also effective (Fig. 5B). Thus, increased tyrosine pho sp horylation of Syk occurs during int egrin-me diat ed cell a dhe sion or subse quen t to integrin ligation by a nti bodies . Immunodepletion experi me nt s revealed that Syk, paxillin, a nd FAK contribute to the overall patterns of tyrosine phosphorylation observed in response to a dhesion or integrin ligation with a nt ibody, but they do not fully a ccou nt for the patterns (data not shown); thus, addit ional unidentified proteins are a lso tyro sine pho sphorylated in respons e to integrin ligation.
A ctivation of Syk Kina se by Engagement of Int egrins-Following st imulation of the cells through integrins, ther e were significant incr ea ses in the activity of the Syk kinase, as indicated by immunocomplex kinase as says (Fig. 6). In nonadherent THP-1 cells , there wa s det ect able kinase act ivity, but the activity increased 4-, 3.5-, and 3-fold, resp ectively, upon cell a dh esion to a fibronectin subst r at u m, ligat ion of {31 inte gri ns with TS2/16 , or with TS2/16 F(ab ' )2; t he radioactivity in ea ch band wa s quantitated using a PhosphorImager. As shown in Fi g. 7, herbimycin A and genis t ein suppressed th e kinase activity of Syk. The concentrat ions of inhibitors required to block Syk activation were similar to tho se required to inhibit the overall pattern of tyro sine phosphorylation and the induction of IL-1/3 message. Thus, in monocytic cells, Syk is an integrinresponsive tyrosine kinase whose activation parallels, and may impinge on, the induction of cytokine messages.
Role of the Cytoskeleton-We have explored the role of the cytoskeleton in the induction of tyrosine phosphorylation and IL-1/3 message expression by use ofthe drug cytochalasin D, an inhibitor of actin filament assembly (38). As seen in Fig. SA, treatment with cytochalasin D inhibits spreading of THP-1 cells on fibronectin but does not affect adhesion. The tyrosine phosphorylation of FAK, and particularly of paxillin, that is induced in THP-1 cells by interaction with a fibronectin substratum is strongly inhibited by cytochalasin D (Fig. 8B). This suggests that some degree of microfilament assembly is required for these phosphorylation events. By contrast, adhesioninduced tyrosine phosphorylation of Syk is not blocked by cytochalasin D treatment; likewise, IL-1/3 message induction is not affected. These observations clearly distinguish between presumptive integrin signaling events that require actin filament assembly and events such as Syk activation that do not have this requirement. They also indicate that extensive cytoskeletal reorganization is not required for integrin-mediated increases in IL-1/3 message levels.

DISCUSSION
Ligation of /31 integrins in peripheral blood monocytes results in a prompt and robust induction of a number of inflammatory mediator genes including IL-1/3, IL-8, IL-6, and tumor necrosis factor, as well as genes for a number of transcription factors (18)(19)(20). Recently, we have demonstrated that /31 integrin ligation in monocytes also results in a burst of tyrosine phosphorylation and that this is necessary for subsequent integrin-mediated gene induction (21). Here, we have shown that similar events, including integrin-mediated tyrosine phosphorylation and increases in inflammatory gene message levels, also occur in THP-1 cells, a monocytic cell line that is more amenable to study of biochemical and molecular events than are primary monocytes. We also show that the transcription factor NF-KB is activated in response to integrin ligation. Further, we have identified some of the proteins that are tyrosine phosphorylated in THP-1 cells subsequent to integrin ligation; these include the focal contact protein paxillin and the FAK and Syk non-receptor tyrosine kinases.
The patterns of tyrosine phosphorylation induced in THP-1 cells by integrin-mediated cell adhesion or by antibody ligation of /31 integrins are similar but distinct. Both types of stimuli result in prominent tyrosine phosphorylation of several components in the 65-75-and the 115-125-kDa ranges. However, as observed in whole cell lysates, integrin-mediated adhesion to fibronectin produces a strong doublet at 65-75 kDa, while antibody ligation results primarily in phosphorylation of the lower molecular mass component of the doublet; likewise, among the tyrosine-phosphorylated proteins at 115-125 kDa, adhesion results in strong phosphorylation of a slowly migrating component (120 kDa), while antibody ligation primarily affects a more rapidly migrating component (115 kDa). The basis for the differences in tyrosine phosphorylation patterns in THP-1 cells stimulated by anti-integrin antibodies and those stimulated by integrin-mediated cell adhesion is unclear at this time. One reasonable possibility is that cell adhesion results in a more complete engagement of the cytoskeleton than does the formation of integrin dimers or multimers triggered by antibodies and the degree of cytoskeletal organization can affect the recruitment of proteins subject to tyrosine phosphorylation.
Despite differences in the overall patterns of tyrosine phosphorylation induced by cell adhesion or by antibody ligation of integrins, it is clear that both of these stimuli can activate NF-KB and can give rise to increased levels of cytokine mes-sages. Furthermore, as we have demonstrated here for THP-1 cells and previously for monocytes (21), inhibitors of tyrosine kinases can block message induction mediated by cell adhesion or by ligation of /31 integrins. These inhibitors also blocked activation of an NF-KB driven promoter-reporter construct in transient transfection assays (data not shown). Thus, it seems highly probable that integrin engagement can trigger activation of tyrosine kinases and protein tyrosine phosphorylation events that are critical for the control of inflammatory mediator genes in monocytic cells. Therefore, it is important to carefully analyze integrin-mediated tyrosine phosphorylation to try to define those events most closely related to the message induction process.
We have examined THP-1 lysates for evidence of integrinmediated tyrosine phosphorylation of several proteins previously identified as being involved in signal transduction cascades or linked to integrin-dependent cytoskeletal reorganization. In THP-1 cells, the focal contact protein paxillin and the cytoplasmic tyrosine kinases FAK and Syk became tyrosine phosphorylated in response to integrin-mediated cell adhesion. The activation ofFAK by integrin ligation, as well as by other stimuli, has been observed in a number of cell types (35,39); paxillin is a substrate for FAK, and its tyrosine phosphorylation seems to parallel the activation of the FAK kinase (23,41). That FAK and paxillin are tyrosine phosphorylated during THP-1 cell adhesion suggests that these molecules may be involved in cytoskeletal reorganization and formation of focal adhesive sites in these cells; a similar role for these proteins has been postulated in fibroblasts (23,42). Since cytochalasin D effectively blocks the tyrosine phosphorylation of FAK and paxillin, some degree of cytoskeletal organization and cell spreading seem to be required for these events.
Ligation of /31 integrins with antibodies did not result in tyrosine phosphorylation of FAK or paxillin but did cause increased tyrosine phosphorylation of Syk, a cytoplasmic tyrosine kinase that contains SH2 domains (24,25). The Syk kinase has been reported to associate with several receptors, including the antigen receptors on B-cells (43), the Fc receptors for IgE (FcERI) (44) and IgG (Fc'YRI,FC'YRIII) (45,46), and granulocyte colony-stimulating factor receptor (26). Syk kinase becomes activated and tyrosine phosphorylated when those receptors bind their respective ligands. A consensus motif has been defined in the cytoplasmic domains of receptors that bind Syk or the related kinase ZAP70 (26). Impaired signal transduction processes have been found in Syk kinase negative cells or in cells that express a Syk mutant lacking kinase activity (43). Since Syk can be activated via Fc receptors, we used F(ab')2 fragments of anti-integrin antibodies to demonstrate that integrin ligation, rather than Fc receptor engagement, was responsible for the increase in Syk enzymatic activity and tyrosine phosphorylation that we observed. In addition, the fact that Syk can be activated by integrin-mediated adhesion to fibronectin substrata also indicates that integrins rather than Fc receptors are involved. Since integrin cytoplasmic domains do not contain any obvious homologies to the Syk binding motif found in other cognate receptors (26), it seems likely that the interaction between integrins and Syk is indirect. In support of this, we have not been able to co-immunoprecipitate Syk and integrins from celllysates (data not shown). Recent studies in platelets have shown that the Syk kinase can be activated via engagement of the allb//33 integrin by fibrinogen (48). Thus, in both monocytic cells and platelets, Syk is an integrin-responsive tyrosine kinase. The activation of Syk, in contrast to the activation of FAK, is not completely dependent on cytoskeletal assembly as observed here as well as in previous studies on platelets (48). Interestingly, recent investigations in rat baso-philic leukemia cells have shown that cross-linking ofthe FCERI receptor can trigger tyrosine phosphorylation of paxillin, while receptor cross-linking in adherent cells can induce tyrosine phosphorylation of both FAK and a stronger phosphorylation of paxillin (47,49). These results may have similar underlying mechanisms to the results presented here, since multi-valent cross-linking of FCERI receptors may cause engagement of the cytoskeleton, setting the stage for FAK and paxillin phosphorylation.
Either integrin-mediated cell adhesion or antibody ligation of integrins can effectively increase IL-ll3 message levels, as well as causing changes in protein tyrosine phosphorylation. However, antibody ligation affected the Syk tyrosine kinase but not FAK or its substrate paxillin. Furthermore, treatment with cytochalasin D, which effectively blocked cell spreading and FAK and paxillin tyrosine phosphorylation, failed to block Syk tyrosine phosphorylation or increases in IL-ll3 message levels. Thus, it seems that FAK is not required for integrin-mediated IL-ll3 message induction in THP-1 cells, consistent with our previous observations on primary monocytes (21). Conversely, our results suggest that Syk may play an important role in the signal transduction process leading from integrin engagement to the induction of inflammatory mediator genes. Thus, the activation ofSyk uniformly precedes and accompanies integrinmediated gene induction, while similar concentrations of tyrosine kinase inhibitors block both Syk autophosphorylation and IL-ll3 message induction. The integrin-mediated changes in tyrosine phosphorylation and Syk activation are transient even during continued cell adhesion or antibody ligation. This suggests that these processes are regulated, perhaps through the agency of protein phosphatases that are induced or activated subsequent to integrin stimulation. Similar events have been associated with the regulation of kinase activities stimulated by mitogens (40).
Current evidence indicates that Syk is an integrin-responsive tyrosine kinase. It also suggests the existence of a signal transduction pathway in monocytic cells that involves the ligation of integrins, activation of the Syk tyrosine kinase, the triggering of downstream events, activation of the NF-KB transcription complex, and the eventual induction of messages for inflammatory mediator genes. Extensive cytoskeletal reorganization does not seem to be required in this pathway, since cytochalasin D treatment had little effect on IL-ll3 message levels. In those types of monocytic cells that do express FAK, such as THP-1, there is a second integrin-mediated signaling pathway that involves FAK activation, and that seems to impact primarily on cytoskeletal organization rather than gene induction.
At this point, a number of issues require further study. It is not certain that activation of the Syk tyrosine kinase is essential for integrin-mediated gene induction, as opposed to simply accompanying the induction process, nor can the participation of additional tyrosine kinases be ruled out. Further experiments will be required to establish a causal role for Syk in integrin signaling. In addition, the downstream events linking changes in tyrosine phosphorylation to alterations of inflammatory gene message levels remain to be defined. In monocytes, it is clear that integrin ligation induces transcriptional activation of the IL-ll3 gene but does not affect message stability.2 In the THP-1 cell system, integrin ligation leads to activation of NF-KB, increased transcription from NF -KB-driven reporter constructs, and increases in IL-ll3 message levels. However, it is not clear at present ifintegrin-mediated changes in IL-ll3 message levels in THP-1 cells are primarily controlled at the transcriptional level or by other means, such as changes in message stability or RNA splicing. If transcriptional activation does occur, the precise role of NF-KB and of other transcription factors remains to be defined. Despite these questions, current observations provide important initial insights into integrin signal transduction processes in monocytic cells and emphasize a key role for specific tyrosine phosphorylation events.