2142.Full.Pdf

E-Selectin-Dependent Signaling Via the Mitogen-Activated Protein Kinase Pathway in Vascular Endothelial Cells

This information is current as Yenya Hu, Jeanne-Marie Kiely, Brian E. Szente, Anthony of September 26, 2021. Rosenzweig and Michael A. Gimbrone, Jr. J Immunol 2000; 165:2142-2148; ; doi: 10.4049/jimmunol.165.4.2142 http://www.jimmunol.org/content/165/4/2142 Downloaded from

References This article cites 47 articles, 28 of which you can access for free at: http://www.jimmunol.org/content/165/4/2142.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 26, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2000 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. E-Selectin-Dependent Signaling Via the Mitogen-Activated Protein Kinase Pathway in Vascular Endothelial Cells1

Yenya Hu,* Jeanne-Marie Kiely,* Brian E. Szente,* Anthony Rosenzweig,† and Michael A. Gimbrone, Jr.2*

E-selectin, a cytokine-inducible adhesion molecule, supports rolling and stable arrest of leukocytes on activated vascular endothelium. Previous studies have suggested that this transmembrane protein can also transduce signals into the endothelial cell. We now demonstrate activation of the mitogen-activated protein kinase (MAPK) signaling cascade in cultured HUVEC in response to E-selectin-dependent leukocyte adhesion and Ab-mediated cross-linking of cell surface E-selectin. Adhesion of increasing numbers of HL60 cells to IL-1␤-activated HUVEC stimulated robust increases in MAPK activity that were abrogated by an E-selectin blocking Ab. Cross-linking of cell surface E-selectin with Abs, as a mimic of multivalent ligand engagement, strongly stimulated MAPK/extracellular signal-related kinase (ERK) kinase (MEK)-dependent MAPK activation and concomitant up- Downloaded from regulation of mRNA for c-fos, an immediate early response gene, whereas Ab cross-linking of HLA class I molecules (present at comparable density) failed to do so. Coimmunoprecipitation documented Ras, Raf-1 and, phospho-MEK complex formation. Unactivated HUVEC transduced with a full-length adenoviral E-selectin construct also exhibited cross-link-induced MAPK activation, macromolecular complex formation, and c-fos up-regulation, whereas HUVEC transduced with a cytoplasmic domain deletion mutant failed to respond. These observations indicate that E-selectin can transduce an activating stimulus via the MAPK cascade into the endothelial cell during leukocyte adhesion. The Journal of Immunology, 2000, 165: 2142–2148. http://www.jimmunol.org/

he selectin family of adhesion molecules, which includes may also play a role in mediating the stable arrest of leukocytes on E-selectin, L-selectin, and P-selectin (1), mediates inter- the luminal surface of inﬂamed microvascular endothelium in the T action of circulating leukocytes with vascular endothe- mouse. lium in various physiological and pathological settings (1). Unique In addition to their function in supporting the physical adhesion among the selectin family molecules, E-selectin typically is not of leukocytes to the luminal surface of the vascular endothelium, detected in unactivated endothelial cells, but is rapidly synthesized recent studies suggest that selectins may also be playing a role in in response to certain cytokines and other pro-inﬂammatory stim- signal transduction during leukocyte-endothelial interactions. For by guest on September 26, 2021 uli, thus making it a marker of the “activated” endothelial pheno- example, our laboratory has shown that leukocyte adhesion to cy- type (2). Along with other selectin family members, E-selectin tokine-activated HUVEC induces clustering of E-selectin mole- exhibits a complex mosaic structure consisting of a large extracel- cules in the vicinity of leukocyte-endothelial cell attachment sites lular portion comprised of an amino-terminal lectin domain, an (8). Leukocyte adhesion to cytokine-activated HUVEC, or the Ab epidermal growth factor domain, and multiple complement regu- induced cross-linking of cell surface E-selectin molecules, results latory repeats, followed by a transmembrane portion and a rela- in a transmembrane linkage of E-selectin to the endothelial cy- tively short (32 aa) cytoplasmic domain (3). P-selectin glycopro- toskeleton via its cytoplasmic domain (8). More recently, Yoshida 3 tein ligand-1 (PSGL-1) can function as a ligand for both E- and et al. (9) have demonstrated that phosphorylation on serine resi- P-selectins (4), but other ligands appear to exist for E-selectin as dues in the cytoplasmic domain of E-selectin is modulated in well (5). E-selectin can support the initial rolling of leukocytes on HUVEC during engagement of E-selectin by leukocytes, Ab cross- activated endothelium as demonstrated in various in vitro and in linking or PSGL-coated beads. Taken together, these data suggest vivo models (3, 6). Milstone et al. (7) have shown that E-selectin that E-selectin can transduce transmembrane signals via its cytoplasmic domain into the endothelial cell. Lorenzon et al. (10) also *Vascular Research Division, Department of Pathology, Brigham and Women’s Hos- have shown that the ligation of either P-selectin or E-selectin with pital, and †Program in Cardiovascular Gene Therapy, Cardiovascular Research Cen- mAbs can induce a transient increase of intracellular ionized cal- ter, Massachusetts General Hospital-East, Boston, MA 02115 cium in endothelial cell, thus further indicating a signaling func- Received for publication January 18, 2000. Accepted for publication May 30, 2000. tion for these vascular selectins. Extensive studies of L-selectin- The costs of publication of this article were defrayed in part by the payment of page dependent signaling in leukocytes also have been undertaken. For charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. example, Ab ligation of L-selectin on the leukocyte surface gen- 1 This work was supported by grants from the National Institutes of Health (P01- erates various transmembrane signals (11, 12), including, in- HL36028 to M.A.G. and HL54202 and AI40970 to A.R.). A.R. is an Established creased intracellular ionized calcium and production of superoxide Investigator of the American Heart Association. ␤ (13), activation of 2 integrin-dependent adhesion (14), and acti- 2 Address correspondence and reprint requests to Dr. Michael A. Gimbrone, Jr., Vas- vation of mitogen-activated protein kinase (MAPK) (15) and c-Jun cular Research Division, Department of Pathology, Brigham and Women’s Hospital, 221 Longwood Avenue, LMRC-401, Boston, MA 02115. E-mail address: N-terminal kinase (JNK) (16) signaling pathways. [email protected] The MAPK cascade (also known as the extracellular signal- 3 Abbreviations used in this paper: PSGL-1, P-selectin glycoprotein ligand-1; MAPK, regulated protein kinase, ERK, pathway) was originally described mitogen-activated protein kinase; ERK, extracellular signal-regulated protein kinase; MEK, MAPK/ERK kinase; GAM, goat anti-murine; WT-E, E-selectin wild type; in cells responding to soluble agonists, such as growth factors and ⌬Cyto-E, E-selectin cytoplasmic deletion. cytokines (17–19). The MAPK cascade consists of a three-kinase

module that includes a MAPK, which is activated by a MAPK/ mixed and centrifuged for 2 min, and the supernatant was analyzed on a ERK kinase (MEK), which in turn is activated by a MEK kinase SDS-PAGE gel. (MEKK). Among the MEKKs, best characterized are the Raf pro- Immunoprecipitation and in vitro kinase assay for MAPK tein isoforms (20). The MAPK pathway can mediate various cel- activity lular responses, including cell motility and shape change, commit- ment to cell cycle or programmed cell death, as well as the MAPK activity was quantified using a kit (p44/42 MAP kinase) from New regulation of multiple genes encoding biologically active products England Biolabs (Beverly, MA), which measures phospho-Elk-1, the phosphorylated product of activated MAPK in a standardized in vitro kinase (21). Activation of MAPK in endothelial cells also has been dem- assay. After treatment, HUVEC were rinsed with ice-cold PBS and lysed onstrated after stimulation by biomechanical force (22) as well as with the kit lysis buffer (20 mM Tris (pH 7.5), 150 mM NaCl, 1 mM cytokine and growth factors (23, 24). Recently, ligand binding by EDTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM ␤-glyc- ␮ integrins or their Ab-mediated cross-linking has been shown to erolphosphate, 1 mM Na3VO4, and 1 g/ml leupeptin) on ice for 5 min. Total cell lysates were sonicated and centrifuged at 14,000 rpm for 15 min; activate MAPK in fibroblasts and endothelial cells (25, 26), and a then supernatants were transferred to new tubes and pre-cleared with pro- similar phenomenon has been observed following Ab cross-linking of tein A/G for1hat4°C. Aliquots (200 ␮l) of these supernatants were the Ig-type adhesion molecule, ICAM-1, in cultured HUVEC (27). incubated overnight at 4°C with the p44/42 MAPK mAb, which specifi- In this study, we have examined the ability of E-selectin-depen- cally recognizes and extracts the phosphorylated (“activated”) species of dent leukocyte adhesion or cross-linking of cell surface E-selectin MAPK. Twenty microliters of protein A/G was incubated with the cell lysates for another hour at 4°C, and the immune complex containing ac- molecules, to initiate outside-in signaling and activation of the tivated MAPK was washed twice with lysis buffer and twice with kinase MAPK pathway in cultured human endothelial cells. Our data sup- buffer (25 mM Tris (pH 7.5), 5 mM ␤-glycerolphosphate, 2 mM DTT, 0.1

␮ Downloaded from port a role for the intact, transmembrane E-selectin molecule as a mM Na3VO4, and 10 mM MgCl2). The pellets were incubated with 48 l ␮ ␮ signal transducer that potentially can inﬂuence multiple events, of kinase buffer, 200 M ATP, and 2 g of Elk-1 protein (the substrate for activated MAPK) at 30°C for 30 min. The reaction then was terminated by including gene regulation, in the endothelial cell during inﬂam- adding 25 ␮lof3ϫ SDS sample buffer. The samples were boiled for 5 min, matory leukocyte recruitment. vortex mixed, and centrifuged for 2 min before Western blotting of the

product, phosphorylated Elk-1 (Elk-1-PO4). Materials and Methods Western blotting analysis

Reagents http://www.jimmunol.org/ ␮ Medium 199, RPMI 1640, and Dulbecco’s PBS (DPBS) were obtained Aliquots (25 l) of immunoprecipitates, prepared as above, were separated from BioWhittaker (Walkersville, MD). FBS was purchased from Life on a 12% SDS-PAGE gel and then transferred to a nylon membrane (Mil- Technologies (Grand Island, NY). Endothelial cell growth factor was ob- lipore, Bedford, MA). Membranes were blocked with 5% nonfat milk in tained from Biomedical Technologies (Stoughton, MA). Paraformaldehyde TTBS (20 mM Tris, 138 mM NaCl, 0.5% Tween 20 (pH 7.6)) for1hat (laboratory grade) was purchased from Fisher Scientific (Springfield, NJ). room temperature and then incubated with various primary Abs (1:1000 Recombinant human IL-1␤ was a gift from Biogen (Cambridge, MA). diluted in blocking buffer), including Ras Ab, MEK 1/2-phospho-specific Biscarboxyethyl-carboxyfluorescein acetoxymethyl ester (BCECF) was Ab (Santa Cruz Biotechnology), or a phospho-Elk-1 polyclonal Ab (1: purchased from Molecular Probes (Eugene, OR). Goat anti-murine 1000) (New England Biolabs) overnight at 4°C. After three washes with (GAM)-IgG immunoglobulin, PD98059, and c-fos oligonucleotide probe TTBS, membranes were incubated with a HRP-conjugated polyclonal goat anti-mouse (or anti-rabbit or anti-rat) Ab (1:1000) (Santa Cruz Biotech- were purchased from Calbiochem (La Jolla, CA). The p44/42 MAP kinase by guest on September 26, 2021 assay kit was obtained from New England Biolabs (Beverly, MA). Protein nology) in TTBS for an additional hour at room temperature, and again A/G-PLUS-Agarose was purchased from Santa Cruz Biotechnology (Santa washed three times in TTBS. The labeled proteins were visualized using an Cruz, CA). enhanced chemilumenscence kit (Amersham, Arlington Heights, IL). Cultured cells Transduction of HUVEC with wild-type and mutant E-selectin via recombinant adenoviral vectors HUVEC were isolated and established in culture as previously described (28). Primary cultures were serially passaged (Ͻ1:3 split ratio) and main- To mediate efficient cell surface expression of E-selectin without activation of tained in Medium 199 buffered with 25 mmol/L HEPES buffer and sup- HUVEC, two replication-defective recombinant type 5 adenoviruses (AdRSV) plemented with 20% FBS, endothelial cell growth factor (25 ␮g/ml), and were used in this study: adenoviral E-selectin wild-type [AdRSV(WT-E)] and porcine intestinal heparin (50 ␮g/ml). For experimental use, subcultured adenoviral E-selectin cytoplasmic deletion mutant [AdRSV(⌬Cyto-E)]. Both (passage 2 or 3) endothelial cells were plated on gelatin-coated 35-mm or constructs use the pJM17 backbone, contain E1/E3 deletions, and were gen- 100-mm tissue culture dishes (Difco Laboratories, Detroit, MI). HL60, a erated as described previously (9). Large-scale production of recombinant vi- human promyelocytic leukocyte cell line, was obtained from the American rus and density gradient purification were performed. High titer (1.5–2.5 ϫ Type Culture Collection (ATCC, Manassas, VA) and grown in RPMI 1640 1012 particles/ml) stocks of each vector were used for these studies. Contam- medium supplemented with 10% FBS, 100 U/ml penicillin, 100 ␮g/ml ination by wild-type adenovirus was excluded by absence of PCR-detectable streptomycin, and 20 mM L-glutamine. JY human lymphocytic cells, E1a sequence in viral stocks. In preliminary experiments, the optimal dose of kindly provided by Dr. T. A. Springer (Center for Blood Research, Boston, adenoviruses to transduce HUVEC was titrated by a fluorescence immuno- MA), were maintained in RPMI 1640 medium supplemented with 10% binding assay to obtain a comparable level of cell surface E-selectin expression FBS, 100 U/ml penicillin, 100 ␮g/ml streptomycin, and 20 mM to that observed on IL-1␤-activated (10 U/ml, 4 h, 37°C) HUVEC (multiplicity L-glutamine. of infection, 64–100 particles per cell). HUVEC (70% confluent) were transduced in M199 containing 10% FCS and used for experimentation 48–72 h Immunoprecipitation for analysis of Ras/Raf-1/phospho-MEK posttransduction, at which time there were no morphologically detectable dif- association ferences between infected and control cultures. After cell surface E-selectin cross-linking, HUVEC were rinsed with ice- Total RNA isolation and Northern blot analysis cold PBS and scraped off the plate in a lysis buffer (20 mM Tris, 5 mM ␮ ␮ MgCl2, 1 mM PMSF, 20 g/ml aproptonin, 10 g/ml leupeptin, 1 mM Total cell RNA was isolated from HUVEC according to the manufacturer’s ␤ Na3VO3, and 20 mM -glycerophosphate). The lysates were then sonicated instructions using RNA STAT-60 (Tel-Test, Friendswood, TX). Ten mil- and centrifuged at 14,000 rpm for 15 min; then the supernatants were ligrams of total RNA was loaded on a 1.2% agarose gel and transferred to transferred to new tubes and pre-cleared with protein A/G for1hat4°C. a nylon membrane. The membrane was subjected to pre-hybridization for Aliquots (200 ␮l) of these supernatants were incubated with a Raf-1 poly- at least4hat42oC. Human c-fos probe was labeled with 5 ␮Ci of clonal Ab at 4°C overnight. Twenty microliters of protein A/G was incu- [␣-32P]ATP and 30 U/␮lof3Ј-terminal deoxynucleotidyl transferase at bated with the cell lysates for another hour at 4°C; then the immune com- 37oC for 1.5 h. The probe was purified using a CHROMA SPIN 10 column plex was washed twice with the lysis buffer and then resuspended in 50 ␮l (Clontech, Palo Alto, CA) and hybridized with a membrane in hybridiza- of lysis buffer. Sample buffer (187.5 mM Tris-HCl (pH 6.8), 6% SDS, 30% tion buffer overnight at 42°C. The membrane was then washed in 2ϫ SSPE glycerol, 150 mM DTT, and 0.3% bromphenol blue) was added to the and 0.1% SDS for 15 min twice at 42°C and 0.2ϫ SSPE and 0.1% SDS for complex and samples were boiled for 5 min. The samples were vortex 30 min at 42°C, then exposed to x-ray film. 2144 E-SELECTIN-MEDIATED MAPK ACTIVATION

Results E-selectin-dependent leukocyte adhesion activates endothelial MAPK To determine whether MAPK is activated as a consequence of leukocyte-endothelial adhesion, HUVEC monolayers were activated with IL-1␤ (10 U/ml, 37°C, 4 h) to stimulate maximal E- selectin cell surface expression, and then incubated with either HL60 cells (a cultured human leukocyte cell line that expresses ligand for E-selectin) (28) or JY cells (another cultured human leukocyte cell line that adheres primarily via LFA-1/ICAM-1) (29). After adhesion under static conditions for 30 min at 37°C, levels of phospho-Elk-1 were measured as an index of MAPK activity. When HL60 cells adhered to IL-1␤-activated HUVEC, MAPK activity increased in proportion to the input concentration of HL60 cells (Fig. 1, lanes 3–5), with robust activation occurring with as few as 2 ϫ 104 cells per well. In contrast, JY cells, which adhered FIGURE 2. A, Fixed HL60 cells fail to show agonist-induced MAPK at comparable density for a given input concentration (data not activity. Live or fixed (2% paraformaldehyde, 4°C, 20 min) HL60 cells shown), showed significantly less MAPK activation in these 4-h IL- were either treated with PMA or vehicle for 15 min, and phospho-Elk-1 (Elk-1-PO ) was measured as an index of MAPK activity. B, Adhesion of Downloaded from 1␤-activated HUVEC monolayers (Fig. 1, lanes 6–8). 4 fixed HL60 cells to IL-1␤-activated HUVEC results in MAPK activation However, the MAPK activity measured in this experimental set- via E-selectin. HUVEC were activated with IL-1␤ (10 U/ml, 37°C, 4 h). ting could be contributed by either the adherent leukocytes or Lane 1, IL-1␤-activated HUVEC alone. Lanes 2–4, Fixed HL60 cells (2 ϫ HUVEC, or both. To investigate whether adhesive interactions be- 106 per well) were incubated with activated HUVEC for the times indi- tween HL60 cells and activated HUVEC result in MAPK activa- cated. Lane 5, An E-selectin adhesion-blocking mAb, 7A9, was preincu- tion in HUVEC, we fixed HL60 cells with 2% paraformaldehyde bated with activated HUVEC (12.5 ␮g/ml, 37°C, 30 min) without the ad- (4°C, 20 min), and then subjected both live and paraformaldehyde- dition of leukocytes. Lanes 6–8, After preincubation of HUVEC http://www.jimmunol.org/ fixed HL60 cells to PMA (a strong stimulus for MAPK activity; 1 monolayers with a mAb, 7A9, fixed HL60 cells (2 ϫ 106 cells per well) ␮M, 15 min) and assayed their lysates for MAPK activity. As seen were added for the times indicated. Phospho-Elk-1 was assayed as an index in Fig. 2A, live HL60 cells exhibited a robust increase in MAPK of MAPK activity and quantified by densitometry; inhibition of leukocyte activity upon PMA stimulation, whereas fixed HL60 cells failed to adhesion was also measured in parallel (see Results). do so. This demonstrated that paraformaldehyde-fixed HL60 cells were not able to activate MAPK even with the input of a strong Ab-mediated cross-linking surface E-selectin activates MAPK stimulus, PMA. We then incubated fixed HL60 with IL-1␤-acti- and this activation is MEK-dependent

vated HUVEC at 37°C for the times indicated in Fig. 2B. The To eliminate the possible contribution of leukocyte cell-associated by guest on September 26, 2021 ␤ adhesion of fixed HL60 to IL-1 -activated HUVEC monolayer stimuli (e.g., growth factors absorbed to the surfaces of the para- activated MAPK in a time-dependent manner (Fig. 2B, lanes 2–4). formaldehyde-fixed HL60 cells), and to mimic the multivalent re- ␮ When an adhesion blocking mAb to E-selectin, 7A9 (12.5 g/ml, ceptor-ligand binding and clustering that occurs at endothelial-leu- ␤ ATCC), was pre-incubated with IL-1 -treated HUVEC, HL60 ad- kocyte surfaces during their adhesive interactions (8), a saturating hesion was inhibited by 56% (average of triplicate samples from amount of a function blocking murine mAb H18/7 (which recog- two separate experiments, 30-min adhesion), as measured by under nizes adhesion supporting epitopes in the extracellular portion of static assay conditions, and MAPK activity was inhibited by 50– the E-selectin molecule) was incubated with IL-1␤-activated 60% (Fig. 2B, lanes 2–4 and 6–8, respectively, as measured by HUVEC and then cross-linked by a GAM-IgG Ab essentially as densitometry). Incubation of fixed HL60 cells with unactivated described previously (8, 9). This cross-linking procedure resulted HUVEC (lacking detectable surface E-selectin expression) re- in marked MAPK activation (Fig. 3, lane 5), while incubation with sulted in negligible adhesion and no detectable MAPK activation either H18/7, or the secondary Ab alone did not do so (Fig. 3, lanes (data not shown). 3 and 4). In contrast, when W6/32, a murine mAb that recognizes surface HLA class I molecules on HUVEC, was similarly cross- linked, MAPK was not significantly activated (Fig. 3, lane 7), although IL-1␤-activated HUVEC exhibit comparable cell surface levels of E-selectin and HLA-class I molecules by fluorescence immunoassay (data not shown). Pre-incubation with a specific MEK inhibitor, PD98059 (20 ␮M), completely inhibited MAPK activation in response to cross-linking of cell surface E-selectin (Fig. 4, lane 3 vs lane 1).

FIGURE 1. Adhesion of leukocytes to IL-1␤-activated HUVEC results MAPK activation is induced by cross-linking cell surface E- in MAPK activation. All HUVEC were serum-starved overnight before the selectin requires its cytoplasmic domain, but does not require experiments, and phospho-Elk-1 was measured as an index of MAPK ac- concomitant cytokine activation of HUVEC tivity (see Materials and Methods). Lane 1, LPS-stimulation of HUVEC To further investigate the signaling role of the cytoplasmic domain (100 ng/ml, 15 min, 37°C), as a positive control for MAPK activation. Lane 2, IL-1␤-activated HUVEC (10 U/ml, 4 h, 37°C). Lanes 3–5 and of E-selectin and the possible inﬂuence of other concomitants of 6–8, Increasing concentrations of either HL60 or JY cells, respectively cytokine activation of HUVEC on E-selectin-dependent MAPK (1ϫϭ2 ϫ 104,10ϫϭ2 ϫ 105, and 100ϫϭ2 ϫ 106 cells per well) were activation, we utilized two adenoviral vectors, AdRSV(WT-E), a incubated with IL-1␤-activated HUVEC for 30 min at 37°C in a standard full length (WT-E) E-selectin, and AdRSV(⌬Cyto-E), a cytoplas- (static) adhesion assay. mic deletion mutant form of E-selectin, to transduce unactivated The Journal of Immunology 2145

FIGURE 3. Cross-linking of cell surface E-selectin, but not HLA class I molecules, activates MAPK in HUVEC. Lane 1, LPS-stimulated HUVEC (positive control for MAPK activity). Lane 2, IL-1␤-activated HUVEC (10 U/ml, 37°C, 4 h; followed by sham incubations at 4°C, 30 min, and 37°C, 30 min) Lane 3, Murine mAb H18/7, which binds to the extracellular domain of E-selectin, was incubated with activated HUVEC (10 ␮g/ml, 4°C, 30 min). Lane 4, GAM-IgG was incubated with activated HUVEC FIGURE 5. A, Unactivated, adeno-transfected HUVEC (WT-E and (sham incubation at 4°C, 30 min; following 1:200, 37°C, 30 min). Lane 5, ⌬Cyto-E) show comparable levels of E-selectin surface expression. Activated HUVEC incubated with murine mAb H18/7 (4°C, 30 min), fol- HUVEC were transfected with adenoviral constructs encoding either full- lowed by GAM-IgG (37°C, 30 min). Lane 6, W6/32, murine mAb to HLA length (WT-E) or cytoplasmic domain deletion mutant (⌬Cyto-E) E-selec- class I molecules, was incubated with activated HUVEC (10 ␮g/ml, 4°C, Downloaded from tin as described in Materials and Methods. Cell surface E-selectin expres- 30 min; followed by sham incubation at 37°C, 30 min). Lane 7, Cell sur- sion as measured by a fluorescence immunobinding assay as described in face HLA class I molecules were cross-linked (W6/32, 4°C, 30 min, fol- Materials and Methods. B, The cytoplasmic domain of E-selectin is nec- lowed by GAM-IgG, 37°C, 30 min). Phospho-Elk-1 (Elk-1-PO ) was as- 4 essary for MAPK activation by Ab cross-linking. Lanes 1–3, HUVEC sayed as an index of MAPK activity. transduced with WT-E-selectin, Lanes 4–6, HUVEC transduced with ⌬Cyto-E-selectin. Lanes 1 and 4, Cells incubated with a murine mAb, cultured HUVEC. Both WT-E- and ⌬Cyto-E-transduced HUVEC H18/7, alone (4°C, 30 min). Lanes 2 and 5, Cells incubated with GAM-IgG http://www.jimmunol.org/ expressed comparable levels of surface E-selectin, as confirmed by alone (37°C, 30 min). Lanes 3 and 6, Cells incubated sequentially with H18/7 and GAM-IgG. Phospho-Elk-1 (Elk-1-PO ) was assayed as an index a fluorescence immunobinding assay, with the same murine E- 4 of MAPK activity. selectin mAb, H18/7, that was used for cross-linking cell surface E-selectin (Fig. 5A). These cell surface levels of E-selectin were similar in magnitude to those observed after standard IL-1␤ stimulation of HUVEC and were not accompanied by any significant the association of both molecules, since it has been well docu- change in other activation markers, such as ICAM-1 or VCAM-1 mented that only GTP-bound Ras is associated with Raf-1 (31–33). ␤ (data not shown), consistent with our previous published results (9, Cell surface E-selectin in IL-1 -activated HUVEC was cross- 30). Cell surface E-selectin in these unactivated HUVECs was then linked by a murine Ab, H18/7, as described in Fig. 3. Raf-1 was by guest on September 26, 2021 cross-linked with a murine E-selectin mAb, H18/7, followed by then immunoprecipitated from the total cell lysates using a poly- GAM-IgG. MAPK activity in HUVEC transduced with WT-E- clonal Ab and this Raf-1-containing immunocomplex was ana- selectin was significantly increased (as was observed in IL-1␤- lyzed by SDS-PAGE followed by immunoblotting using specific activated HUVEC), while HUVEC transduced with ⌬Cyto-E-se- Abs against Ras and phospho-MEK. Cross-linking cell surface E- lectin failed to generate any MAPK activation signal (Fig. 5B). selectin resulted in increased amounts of Ras in the Raf-1-containing immunocomplex (Fig. 6A). Similarly, probing with an anti- Cross-linking cell surface E-selectin results in the formation of phospho-MEK-specific Ab also revealed increased amounts of this a Ras/Raf-1/phospho-MEK macromolecular complex and that is component in the Raf-1-containing complex (Fig. 6A). Re-probing cytoplasmic domain-dependent the same blots with a Raf-1 Ab, after stripping, confirmed comparable immunoprecipitation of Raf-1 from both control and cross- To investigate the signaling pathway upstream of MAPK activa- linked samples (Fig. 6A). A reciprocal immunoprecipitation, in tion, we examined the activation of Ras and Raf-1 by determining which phospho-MEK was immunoprecipitated from total cell ly- sate and the resultant immunocomplex analyzed using Abs against Ras and Raf-1, revealed increased association of Ras and Raf-1 with phospho-MEK following E-selectin cross-linking (data not shown). To investigate the role of the cytoplasmic domain of E- selectin in this macromolecular complex formation, HUVEC were transduced with either WT-E-selectin or ⌬Cyto-E-selectin adenoviral constructs, as described in Materials and Methods. Cell surface E-selectin molecules were cross-linked by a murine mAb, H18/7, followed by a GAM-IgG. As shown in Fig. 6B, cross- linking WT-E-selectin (left lanes in each panel), but not ⌬Cyto- FIGURE 4. MAPK activation induced by cross-linking cell surface E- E-selectin (right lanes in each panel), resulted in increased asso- selectin is MEK-dependent. Lane 1, Cross-linking of cell surface E-selectin ciation of Raf-1 with Ras and phospho-MEK. in IL-1␤-activated HUVEC (10 U/ml, 4 h, 37°C) via sequential incubation with a murine anti-E-selectin mAb, H18/7, and a GAM-IgG (4°C, 30 min Cross-linking surface E-selectin up-regulates c-fos at mRNA and 37°C, 30 min, respectively). Lane 2, A specific MEK inhibitor, level and the cytoplasmic domain of E-selectin is required for ␮ ␤ PD98059 (20 M) incubated with IL-1 -activated HUVEC (37°C, 45 this up-regulation min). Lane 3, PD98059 (20 ␮M) was pre-incubated (37°C, 45 min) with activated HUVEC before the cross-linking procedure. Phospho-Elk-1 (Elk- To determine whether E-selectin-mediated MAPK signaling can

1-PO4) was assayed as an index of MAPK activity. result in gene regulation, total RNA was isolated from HUVEC 2146 E-SELECTIN-MEDIATED MAPK ACTIVATION

FIGURE 6. A, Cross-linking cell surface E-selectin results in the formation of a macromolecular complex containing Ras, Raf-1, and phospho-

MEK. Total cell lysates prepared from HUVEC with and without cross- Downloaded from linking of cell surface E-selectin were subjected to immunoprecipitation (IP) with a Raf-1 polyclonal Ab, and the resulting immunocomplex was analyzed by SDS-PAGE gel and immunoblotting (IB) with appropriate FIGURE 7. A, Cross-linking cell surface E-selectin up-regulates c-fos Abs to detect the presence of Ras, Raf-1, and phospho-MEK. Left lane in mRNA in an MAPK-dependent manner. Lanes 1–7, HUVEC was activated ␤ ␤ each panel, IL-1␤-activated HUVEC (10 U/ml, 37°C, 4 h) alone. Right with IL-1 (10 U/ml, 37°C, 4 h). Lane 1, IL-1 -activated HUVEC alone. ␮ ␤ lane in each panel, IL-1␤-treated HUVEC following sequential incubation Lane 2, H18/7 (10 g/ml) was incubated with IL-1 -activated HUVEC with murine mAb H18/7 (4°C, 30 min) and GAM-IgG (37°C, 30 min) to (4°C, 30 min). Lane 3, GAM-IgG (37°C, 30 min). Lane 4, Cell surface http://www.jimmunol.org/ cross-link cell surface E-selectin. Note that the association of Ras and E-selectin molecules were cross-linked by H18/7 followed by GAM-IgG. ␮ phospho-MEK with Raf-1 is strongly enhanced by E-selectin cross-linking, Lane 5, PD98059 was incubated with HUVEC (20 M, 37°C, 45 min). and that comparable amounts of Raf-1 are present in HUVEC lysates be- Lane 6, HLA class I molecules were cross-linked using W6/32 followed by fore and after E-selectin cross-linking. B, The cytoplasmic domain of E- GAM-IgG. Lane 7, PD98059 was incubated with HUVEC before the cell selectin is required for formation of the Ras, Raf-1, and phospho-MEK surface E-selectin cross-linking. Hybridization of a GAPDH probe to the macromolecular complex. Cell surface E-selectin molecules were cross- same blot reveals comparable loading of total RNA in each lane. B, The linked by sequential incubation with murine mAb, H18/7 (4°C, 30 min) cytoplasmic domain of E-selectin is required for the E-selectin-dependent and GAM-IgG (37°C, 30 min) on HUVEC transduced with either WT-E- up-regulation of c-fos at mRNA level. Total RNA was isolated and ana- ⌬ lyzed. Lanes 1 and 3, HUVEC transduced with either WT-E-selectin or

selectin or Cyto-E-selectin adenoviral constructs. Total cell lysates were by guest on September 26, 2021 ⌬ subjected to immunoprecipitation (IP) using a Raf-1 polyclonal Ab, and the Cyto-E-selectin, respectively. Lanes 2 and 4, Cell surface E-selectin mol- Raf-1-containing immunocomplex was analyzed by SDS-PAGE and im- ecules were Ab cross-linked (as in A) on HUVEC transduced either with ⌬ munoblotted (IB) with either an anti-Ras Ab or an anti-phospho-MEK Ab. WT-E-selectin or with Cyto-E-selectin, respectively. Left lanes on each panel, WT-E-selectin transduced HUVEC; Right lanes on each panel, ⌬Cyto-E-selectin-transduced HUVEC. extracellular domain of E-selectin, thus indicating its direct involvement in this signaling event. Interestingly, adhesion of JY that were treated with IL-1␤ (10 U/ml, 37°C, 4 h) and then surface cells, another human leukocyte cell line, that adheres to a compa- cross-linked with either the anti-E-selectin murine mAb, H18/7, or rable extent to 4-h IL-1␤-activated HUVEC but largely via an the anti-HLA class I murine mAb, W6/32. Northern blotting was ICAM-1/LFA-1 interaction (29), does not result in robust MAPK conducted to measure the steady-state mRNA levels of c-fos,an activation (Fig. 1). However, this does not rule out a possible sig- immediate early response gene that can be regulated via the naling role for ICAM-1 during leukocyte-endothelial adhesion, as MAPK pathway (34, 35). As seen in Fig. 7A, the c-fos mRNA suggested in previous studies (27, 36). Nonetheless, our data do level was markedly increased only when cell surface E-selectin suggest that E-selectin may play a more dominant role in MAPK was cross-linked (Fig. 7A, lane 4). The MEK inhibitor, PD98059, activation during the early phases of anti-inflammatory response, completely inhibited this up-regulation (Fig. 7A, lane 7). Cross- as mimicked by short-term cytokine stimulation in this in vitro linking of cell surface HLA class I molecules using the mAb, model system. W6/32 did not result in c-fos up-regulation (Fig. 7A, lane 6). In our previous studies, the use of Ab-mediated cross-linking to Cross-linking WT-E-selectin (Fig. 7B, lanes 1 and 2), but not mimic the clustering of cell surface E-selectin that occurs during ⌬Cyto-E-selectin (Fig. 7B, lanes 3 and 4), up-regulated c-fos at leukocyte-endothelial interaction resulted in cytoskeletal associa- mRNA level. tion and dephosphorylation events that were dependent on the cytoplasmic domain of this transmembrane protein (8, 9). As seen in Discussion Fig. 3, when surface E-selectin was cross-linked by the same Ab In this report we show that E-selectin, an inducible adhesion mol- treatment, MAPK activity increased dramatically. In contrast, ecule expressed on the surface of activated endothelial cells, can cross-linking another endothelial surface molecule, the HLA class transduce signals across the cell membrane to activate the MEK- I heterodimer, present at comparable density on the surface of dependent MAPK cascade. Adhesion of both live and fixed (met- IL-1␤-treated HUVEC, did not generate a comparable level of abolically inert) HL60 cells to IL-1␤-activated HUVEC resulted in MAPK activation. Thus, nonspecific perturbation of the cell mem- the activation of MAPK in a dose- and time-dependent fashion brane by Ab cross-linking does not appear to be responsible for the (Figs. 1 and 2). MAPK activation was proportionally inhibited observed MAPK activation. Further, since cross-linking surface with a specific adhesion-blocking mAb, 7A9, that binds to the E-selectin via a non-adhesion-blocking mAb, H4/18 (which also The Journal of Immunology 2147 interacts with the extracellular domain of E-selectin), comparably In addition to its involvement in various basic aspects of cell activated MAPK, it appears that receptor-clustering per se is a biology (e.g., cell motility, cell shape, cell cycle, apoptosis), sig- sufficient stimulus. In the preliminary experiments, binding of naling via the MAPK pathway has been shown to influence the beads coated with an E-selectin ligand, PSGL-1, also induced regulation of genes encoding a broad spectrum of biologically ac- MAPK activation (Y. Hu, unpublished observation). Taken to- tive products, including chemokines (34, 44) and adhesion mole- gether, we interpret these data to indicate that clustering of E- cules (25, 26, 35, 45, 46). c-fos is an example of an immediate selectin molecules at the cell surface, induced by leukocyte adhe- early gene that encodes a transcription factor that is involved in the sion, or Ab- or ligand-induced cross-linking, can act as a sufficient transcriptional regulation of multiple genes. MAPK activation can stimulus for activation of the MAPK pathway. However, our cur- result in c-fos up-regulation (34, 35, 47, 48). In our experiments, rent studies do not establish that E-selectin clustering, as occurs in c-fos mRNA was up-regulated in a MAPK-dependent manner the context of leukocyte adhesion to the surface of an activated when cell surface E-selectin molecules, but not surface HLA class endothelial cell, is the sole mechanism for MAPK pathway acti- I molecules, were cross-linked (Fig. 7A). In contrast, the cytoplas- vation. Further, it is also possible that E-selectin clustering may mic domain deletion mutant of E-selectin failed to up-regulate c- result in the activation, in parallel, of signaling pathways (e.g., fos expression after cross-linking (Fig. 7B), which again suggests the direct involvement of this portion of the E-selectin molecule in c-Jun NH2-terminal kinase/stress-activated protein kinase) in addition to the MAPK cascade. MAPK signaling. We are currently characterizing the temporal Although selectins share highly homologous mosaic domains in pattern of expression of multiple endothelial genes, associated with their extracellular portions, their respective cytoplasmic domains E-selectin-dependent MAPK activation via a transcriptional pro- are distinct (1, 3) and these divergent structures appear to support filing strategy in an effort to define this aspect of leukocyte adhe- Downloaded from different functions. For example, the cytoplasmic domain of L- sion-induced phenotypic modulation. selectin plays a critical role in neutrophil rolling in vivo at sites of In summary, we have demonstrated that E-selectin can act as a inflammation, and in the binding of lymphocytes to high endothe- transmembrane signal transducer, activating the MAPK cascade lial venules of peripheral lymph node tissue (37). In contrast, the and resulting in the up-regulation of the immediate early response deletion of the cytoplasmic domain of P-selectin does not affect gene, c-fos, which is itself further implicated in the transcriptional leukocyte adhesion (38). However, the cytoplasmic domain of P- control of various pro-inflammatory genes. E-selectin-dependent http://www.jimmunol.org/ leukocyte adhesion-induced modulation of endothelial phenotype selectin does appear to play a critical role in the intracellular sort- may have important implications for the evolution of the inflam- ing of P-selectin to storage granules in endothelial cells and plate- matory process. Further studies of the molecular mechanisms link- lets (38). HUVEC transduced with a cytoplasmic domain deletion ing leukocyte adhesion-dependent rearrangement of E-selectin at mutant E-selectin show a comparable level of cell surface expres- the cell surface to intracellular cascade signaling cascades, such as sion of E-selectin protein as those transduced with WT-E-selectin, the MAPK pathway, may provide valuable insights into the or- and also support comparable HL60 cell adhesion under nonstatic chestration of the inflammatory response at the level of the vas- adhesion assay conditions (9). However, deletion of the cytoplas- cular endothelial lining.

mic domain of E-selectin does disrupt adhesion-induced cytoskel- by guest on September 26, 2021 etal association and dephosphorylation (8, 9), thus suggesting that Acknowledgments this portion of the molecule is playing an important role in signal- We gratefully acknowledge the expert assistance of Kay Case and William ing. In the current study, we now demonstrate that the deletion of Atkinson in cell culture. We also thank Drs. Andrew Connolly and Francis the cytoplasmic domain of E-selectin results in the loss of the W. Luscinskas for helpful discussions. robust MAPK activation that is induced by cross-linking of cell surface E-selectin (Fig. 5B). References We observed that Ras and Raf-1 form an immunoprecitable 1. Kansas, G. S. 1996. Selectins and their ligands: current concepts and controver- complex when cell surface E-selectin is cross-linked (Fig. 6A). sies. Blood 88:3259. 2. Bevilacqua, M. P., J. S. Pober, D. L. Mendrick, R. S. Cotran, and This association appeared within 10 min and was sustained for at M. A. Gimbrone, Jr. 1987. Identiﬁcation of an inducible endothelial-leukocyte least 30 min after cross-linking, which is consistent with typical adhesion molecule. Proc. Natl. Acad. Sci. USA 84:9238. time course of stimulation of Ras (33). It has been shown that, 3. Bevilacqua, M. P., and R. M. Nelson. 1993. Selectins. J. Clin. Invest. 91:379. 4. McEver, R. P., and R. D. Cummings. 1997. Role of PSGL-1 binding to selectins upon receptor activation, Raf-1 is recruited to the plasma mem- in leukocyte recruitment. J. Clin. Invest. 100:S97. brane and becomes associated selectively with GTP-bound Ras 5. Yang, J., B. C. Furie, and B. Furie. 1999. The biology of P-selectin glycoprotein ligand-1: its role as a selectin counterreceptor in leukocyte-endothelial and leu- (33, 39). Numerous reports also have shown that Ras and Raf-1 kocyte-platelet interaction. Thromb. Haemostasis 81:1. can form a signaling complex with MEK, in which Raf-1 mediates 6. Jung, U., K. E. Norman, K. Scharffetter-Kochanek, A. L. Beaudet, and K. Ley. phosphorylation of MEK on serine residues (40–43). Consistent 1998. Transit time of leukocytes rolling through venules controls cytokine-induced inﬂammatory cell recruitment in vivo. J. Clin. Invest. 102:1526. with this, we found that surface E-selectin cross-linking resulted in 7. Milstone, D. S., D. Fukumura, R. C. Padgett, P. E. O’Donnell, V. M. Davis, increased amounts of phosphorylated MEK in the Raf-1/Ras com- O. J. Benavidez, W. L. Monsky, R. J. Melder, R. K. Jain, and M. A. Gimbrone, Jr. 1998. Mice lacking E-selectin show normal numbers of rolling leukocyte but plex. Thus, a functional macromolecular complex (Ras/Raf-1/ reduced leukocyte stable adhesion to cytokine-activated microvascular endothe- phosphorylated MEK) forms as a consequence of cell surface E- lium. Microcirculation 5:153. selectin cross-linking. Unactivated HUVEC transduced with a 8. Yoshida, M., W. F. Westlin, N. Wang, D. E. Ingber, A. Rosenzweig, N. Resnick, and M. A. Gimbrone, Jr. 1996. Leukocyte adhesion to vascular endothelium cytoplasmic deletion mutant (⌬Cyto-E) failed to generate this Ras/ induces E-selectin linkage to the actin cytoskeleton. J. Cell Biol. 133:445. Raf-1/phospho-MEK macromolecular complex upon cell surface 9. Yoshida, M., B. E. Szente, J. M. Kiely, A. Rosenzweig, and M. A. Gimbrone, Jr. 1998. Phosphorylation of the cytoplasmic domain of E-selectin is regulated dur- E-selectin cross-linking (Fig. 6B), thus indicating that the cyto- ing leukocyte-endothelial adhesion. J. Immunol. 161:933. plasmic domain may play an important role in the activation of 10. Lorenzon, P., E. Vecile, E. Nardon, E. Ferrero, J. M. Harlan, F. Tedesco, and Ras. It has been previously shown that Ras can become indirectly A. Dobrina. 1998. Endothelial cell E- and P-selectin and vascular cell adhesion molecule-1 function as signaling receptors. J. Cell Biol. 142:1381. associated with L-selectin via adapter proteins during Ab-mediated 11. Laudanna, C., G. Constantin, P. Baron, E. Scarpini, G. Scarlato, G. Cabrini, cross-linking (15). We are currently exploring the exact mecha- C. Dechecchi, F. Rossi, M. A. Cassatella, and G. Berton. 1994. Sulfatides trigger increase of cytosolic free calcium and enhanced expression of tumor necrosis nism by which the cytoplasmic domain of E-selectin interacts with factor-␣ and interleukin-8 mRNA in human neutrophils: evidence for the role of Ras in endothelial cells. L-selectin as a signaling molecule. J. Biol. Chem. 269:4021. 2148 E-SELECTIN-MEDIATED MAPK ACTIVATION

12. Steeber, D. A., P. Engel, A. S. Miller, M. P. Sheetz, and T. F. Tedder. 1997. lial cells under simulated physiological ﬂow conditions in vitro. Circ. Res. 79: Ligation of L-selectin through conserved regions within the lectin domain acti- 1205. vates signal transduction pathways and integrin function in human, mouse, and 31. Koide, H., T. Satoh, M. Nakafuku, and Y. Kaziro. 1993. GTP-dependent asso- rat leukocytes. J. Immunol. 159:952. ciation of Raf-1 with Ha-Ras: identiﬁcation of Raf as a target downstream of Ras 13. Crockett-Torabi, E., and J. C. Fantone. 1997. L-selectin stimulation of canine in mammalian cells. Proc. Natl. Acad. Sci. USA 90:8683. neutrophil initiates calcium signal secondary to tyrosine kinase activation. 32. Marshall, M. 1995. Interactions between Ras and Raf: key regulatory proteins in Am. J. Physiol. 272:H1302. cellular transformation. Mol. Reprod. Dev. 42:493. 14. Simon, S. I., A. R. Burns, A. D. Taylor, P. K. Gopalan, E. B. Lynam, L. A. Sklar, 33. Denhardt, D. T. 1996. Signal-transducing protein phosphorylation cascades me- and C. W. Smith. 1995. L-selectin (CD62L) cross-linking signals neutrophil ad- diated by Ras/Rho proteins in the mammalian cell: the potential for multiplex ␤ hesive functions via the Mac-1 (CD11b/CD18) 2 integrin. J. Immunol. 155: signalling. Biochem. J. 318:729. 1502. 34. Jalali, S., Y. S. Li, M. Sotoudeh, S. Yuan, S. Li, S. Chien, and J. Y. Shyy. 1998. 15. Brenner, B., E. Gulbins, K. Schlottmann, U. Koppenhoefer, G. L. Busch, Shear stress activates p60src-Ras-MAPK signaling pathways in vascular endo- B. Walzog, M. Steinhausen, K. M. Coggeshall, O. Linderkamp, and F. Lang. thelial cells. Arterioscler. Thromb. Vasc. Biol. 18:227. lck 1996. L-selectin activates the Ras pathway via the tyrosine kinase p56 . Proc. 35. Huttunen, P., T. Hyypia, P. Vihinen, L. Nissinen, and J. Heino. 1998. Echovirus Natl. Acad. Sci. USA 93:15376. 1 infection induces both stress- and growth-activated mitogen-activated protein 16. Brenner, B., S. Weinmann, H. Grassme, F. Lang, O. Linderkamp, and E. Gulbins. kinase pathways and regulates the transcription of cellular immediate-early 1997. L-selectin activates JNK via src-like tyrosine kinases and the small G- genes. Virology 250:85. protein Rac. Immunology 92:214. 36. Durieu-Trautmann, O., N. Chaverot, S. Cazaubon, A. D. Strosberg, and 17. Xing, J., J. M. Kornhauser, Z. Xia, E. A. Thiele, and M. E. Greenberg. 1998. P. O. Couraud. 1994. Intercellular adhesion molecule 1 activation induces ty- Nerve growth factor activates extracellular signal-regulated kinase and p38 mi- rosine phosphorylation of the cytoskeleton-associated protein cortactin in brain togen-activated protein kinase pathways to stimulate CREB serine 133 phosphor- microvessel endothelial cells. J. Biol. Chem. 269:12536. ylation. Mol. Cell. Biol. 18:1946. 37. Kansas, G. S., K. Ley, J. M. Munro, and T. F. Tedder. 1993. Regulation of 18. Stancato, L. F., M. Sakatsume, M. David, P. Dent, F. Dong, E. F. Petricoin, ␤ leukocyte rolling and adhesion to high endothelial venules through the cytoplas- J. J. Krolewski, O. Silvennoinen, P. Saharinen, J. Pierce, et al. 1997. Interferon mic domain of L-selectin. J. Exp. Med. 177:833. and oncostatin M activate Raf-1 and mitogen-activated protein kinase through a

38. Hartwell, D. W., T. N. Mayadas, G. Berger, P. S. Frenette, H. Rayburn, Downloaded from JAK1-dependent pathway. Mol. Cell. Biol. 17:3833. R. O. Hynes, and D. D. Wagner. 1998. Role of P-selectin cytoplasmic domain in 19. May, M. J., C. P. Wheeler-Jones, R. A. Houliston, and J. D. Pearson. 1998. granular targeting in vivo and in early inﬂammatory responses. J. Cell Biol. Activation of p42mapk in human umbilical vein endothelial cells by interleu- 143:1129. kin-1␣ and tumor necrosis factor-␣. Am. J. Physiol. 274:C789. 39. Overbeck, A. F., T. R. Brtva, A. D. Cox, S. M. Graham, S. H. Huff, 20. Robinson, M. J., and M. H. Cobb. 1997. Mitogen-activated protein kinase path- R. Khosravi-Far, L. A. Quilliam, P. A. Solski, and C. J. Der. 1995. Guanine ways. Curr. Opin. Cell Biol. 9:180. nucleotide exchange factors: activators of Ras superfamily proteins. Mol. Reprod. 21. Yamada, K. M., and B. Geiger. 1997. Molecular interactions in cell adhesion Dev. 42:468. complexes. Curr. Opin. Cell Biol. 8:657. 40. Van Aelst, L., M. Barr, S. Marcus, A. Polverino, and M. Wigler. 1993. Complex

22. Traub, O., and B. C. Berk. 1998. Laminar shear stress: mechanisms by which http://www.jimmunol.org/ formation between RAS and RAF and other protein kinases. Proc. Natl. Acad. endothelial cells transduce an atheroprotective force. Arterioscler. Thromb. Vasc. Sci. USA 90:6213. Biol. 18:677. 23. Milanini, J., F. Vinals, J. Pouyssegur, and G. Pages. 1998. p42/p44 MAP kinase 41. Wartmann, M., and R. J. Davis. 1994. The native structure of the activated Raf module plays a key role in the transcriptional regulation of the vascular endo- protein kinase is a membrane-bound multi-subunit complex. J. Biol. Chem. 269: thelial growth factor gene in fibroblasts. J. Biol. Chem. 273:18165. 6695. 24. Doanes, A. M., D. D. Hegland, R. Sethi, I. Kovesdi, J. T. Bruder, and T. Finkel. 42. Moodie, S. A. W., Berthe M. Weber, Michael J. Wolfman Alan. 1993. Com- 1999. VEGF stimulates MAPK through a pathway that is unique for receptor plexes of Ras:GTP with Raf-1 and mitogen-activated protein kinase kinase. Sci- tyrosine kinases. Biochem. Biophys. Res. Commun. 255:545. ence 260:1658. 25. Schlaepfer, D. D., S. K. Hanks, T. Hunter, and P. van der Geer. 1994. Integrin- 43. Huang, W. A., Alessandro Crews, Craig M. Erikson, R.L. 1993. Raf-1 forms a mediated signal transduction linked to Ras pathway by GRB2 binding to focal stable complex with Mek1 and activates Mek1 by serine phosphorylation. Proc. adhesion kinase. Nature 372:786. Natl. Acad. Sci. USA 90:10947. 44. Fischer, T. A., K. Singh, D. S. O’Hara, D. M. Kaye, and R. A. Kelly. 1998. Role 26. Short, S. M., G. A. Talbott, and R. L. Juliano. 1998. Integrin-mediated signaling by guest on September 26, 2021 events in human endothelial cells. Mol. Biol. Cell 9:1969. of AT1 and AT2 receptors in regulation of MAPKs and MKP-1 by ANG II in 27. Lawson, C., M. Ainsworth, M. Yacoub, and M. Rose. 1999. Ligation of ICAM-1 adult cardiac myocytes. Am. J. Physiol. 275:H906. on endothelial cells leads to expression of VCAM-1 via a nuclear factor-␬B- 45. Herrera, R., S. Hubbell, S. Decker, and L. Petruzzelli. 1998. A role for the MEK/ independent mechanism. J. Immunol. 162:2990. MAPK pathway in PMA-induced cell cycle arrest: modulation of megakaryocytic 28. Bevilacqua, M. P., J. S. Pober, M. E. Wheeler, R. S. Cotran, and M. A. Gimbrone, differentiation of K562 cells. Exp. Cell Res. 238:407. Jr. 1985. Interleukin 1 acts on cultured human vascular endothelium to increase 46. Ahmad, M., P. Theofanidis, and R. M. Medford. 1998. Role of activating pro- the adhesion of polymorphonuclear leukocytes, monocytes, and related leukocyte tein-1 in the regulation of the vascular cell adhesion molecule-1 gene expression cell lines. J. Clin. Invest. 76:2003. by tumor necrosis factor-␣. J. Biol. Chem. 273:4616. 29. Rothlein, R., and T. A. Springer. 1986. The requirement for lymphycyte function 47. Zutter, M. M., A. D. Painter, and X. Yang. 1999. The megakaryocyte/platelet- ␣ ␤ associated-antigen 1 in homotypic leukocyte adhesion stimulated by phorbol es- specific enhancer of the 2 1 integrin gene: two tandem AP1 sites and the mi- ter. J. Exp. Med. 163:1132. togen-activated protein kinase signaling cascade. Blood 93:1600. 30. Gerszten, R. E., F. W. Luscinskas, H. T. Ding, D. A. Dichek, L. M. Stoolman, 48. Cavigelli, M., F. Dolfi, F. X. Claret, and M. Karin. 1995. Induction of c-fos M. A. Gimbrone, Jr., and A. Rosenzweig. 1996. Adhesion of memory lympho- expression through JNK-mediated TCF/Elk-1 phosphorylation. EMBO J. cytes to vascular cell adhesion molecule-1-transduced human vascular endothe- 14:5957.