Cytokine signaling in endothelial cells

Endothelial cells mediate the inflammatory response by coordinating localization and activation of leukocytes via a concerted expression of adhesion molecules and cytokines. Adhesion molecule and cytokine gene expression occurs following activation of specific signaling pathways by endothelial cell agonists. The MAP kinase signaling pathways consist of a cascade of kinases that phosphorylate transcription factors to activate specific genes. The major MAP kinase pathways are the 'ERK' or 'growth' pathway and the 'stress' pathway. NF-?B, another mediator of endothelial gene expression, is normally sequestered by the I?B in the cytoplasm. However, certain stimuli cause degradation of I?B that results in NF-?B nuclear translocation. In the nucleus NF-?B activates transcription from NF-?B dependent promoters of adhesion molecule and cytokine genes. The major aim of this study was to describe signaling mechanisms involved in the induction of adhesion molecules and cytokines by the use of receptor and nonreceptor endothelial cell agonists. The nonreceptor agonists brown recluse spider venom and bacterial sphingomyelinase induced a pattern of gene expression distinct from tumor necrosis factor (TNF). Since TNF activation of NF-?B and MAP kinase was theorized to be downstream of sphingomyelin breakdown due to a TNF receptor associated sphingomyelinase, the results with nonreceptor sphingomyelin breakdown were intriguing. On elucidation of signaling pathways, TNF sends at least two signals. Sphingomyelin breakdown to ceramide activates the ERK MAP kinase pathway, and ceramide-independent activation of stress MAP kinases and NF-?B nuclear translocation. Another novel receptor agonist oncostatin M (OSM) was found to induce a pattern of gene expression distinct from both TNF and bacterial sphingomyelinase. At low concentrations, OSM caused potent ERK activation, but at higher concentrations it caused NF?B nuclear translocation in addition to ERK activation. This unmasked NF-?B-dependent and -independent mechanisms of endothelial cell adhesion molecule and cytokine expression. Furthermore, the above agonists did not induce the same characteristic responses in gene induction and signaling in transformed endothelial cells, showing that alteration in signaling pathways associated with cell transformation alters gene expression. In conclusion, integration of different signals at the promoters of adhesion molecule and cytokine genes mediates a coordinated inflammatory response by endothelial cells.

CYTOKINE SIGNALING IN ENDOTHELIAL CELLS by Vijayanand R. Modur A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Experimental Pathology Department of Pathology The University of Utah June 1996 Copyright © Vijayanand R. Modur 1996 All Rights Reserved THE UNIVERSITY OF UTAH GRADUATE SCHOOL SUPERVISORY COMMITTEE APPROVAL ofa~nationsubnrittedby Vijayanand R. Modur This dissenation has been read by each member of the following supervisory committee and by majority vote has been found to be satisfactory. Stephen M. Prescott THE UNIVERSITY OF UTAH GRADUATE SCHOOL FINAL READING APPROVAL To the Graduate Council of the University of Utah: I have read the dissertation of Vi; ayanand R. Modur in its fmal form and have found that (1) its format, citationsy and bibliographic style are consistent and acceptable; (2) its illustrative materials including figures, tables, and charts are in place; and (3) the final manuscript is satisfactory to the supervisory committee and is ready for submission to The Graduate School. Chair, Supervisory Committee Approved for the Major Department ~~~ Raymo A. Daynes Chair1Dean Approved for the Graduate Council Dean of The Graduate School ABSTRACT Endothelial cells mediate the inflammatory response by coordinating localization and activation of leukocytes via a concerted expression of adhesion molecules and cytokines. Adhesion molecule and cytokine gene expression occurs following activation of specific signaling pathways by endothelial cell agonists. The MAP kinase signaling pathways consist of a cascade of kinases that phosphorylate transcription factors to activate specific genes. The major MAP kinase pathways are the "ERK" or "growth" pathway and the "stress" pathway. NF-KB, another mediator of endothelial gene expression, is normally sequestered by the IKB in the cytoplasm. However, certain stimuli cause degradation of I KB that results in NF-KB nuclear translocation. In the nucleus NF-KB activates transcription from NF-KB dependent promoters of adhesion molecule and cytokine genes. The major aim of this study was to describe signaling mechanisms involved in the induction of adhesion molecules and cytokines by the use of receptor and nonreceptor endothelial cell agonists. The nonreceptor agonists brown recluse spider venom and bacterial sphingomyelinase induced a pattern of gene expression distinct from tumor necrosis factor (TNF). Since TNF activation of NF-KB and MAP kinase was theorized to be downstream of sphingomyelin breakdown due to a TNF receptor associated sphingomyelinase, the results with nonreceptor sphingomyelin breakdown were intriguing. On elucidation of signaling pathways, TNF sends at least two signals. Sphingomyelin breakdown to cerami de activates the ERK MAP kinase pathway, and ceramide-independent activation of stress MAP kinases and NF-KB nuclear translocation. Another novel receptor agonist oncostatin M (OSM) was found to induce a pattern of gene expression distinct from both TNF and bacterial sphingomyelinase. At low concentrations, OSM caused potent ERK activation, but at higher concentrations it caused NF-KB nuclear translocation in addition to ERK activation. This unmasked NF-KB-dependent and -independent mechanisms of endothelial cell adhesion molecule and cytokine expression. Furthermore, the above agonists did not induce the same characteristic responses in gene induction and signaling in transformed endothelial cells, showing that alteration in signaling pathways associated with cell transformation alters gene expression. In conclusion, integration of different signals at the promoters of adhesion molecule and cytokine genes mediates a coordinated inflammatory response by endothelial cells. v In the memory of my brother, Sanjay ... TABLE OF CONTENTS ABSTRACT .................................................................................................................... iv ACKNOWLEDGMENTS ............................................................................................. ix Chapter I. INTRODUCTION............ ...................... ................. ....... .... ........ .... .................... 1 Physiological basis of adhesion........... ......... ...... ........ ...... ......................... 1 Adhesion molecules........... ............. ............ ............. ........... .......... ............. 4 Cytokines expressed by endothelial cells..... .............. .......... ......... .......... 5 Signaling...... ...... ....... ...... ..... .............. ............ ............ ......................... ........... 6 References ...................................................................................................... 12 II. MATERIALS AND METHODS ...................................................................... 17 Materials ........................................................................................................ 17 Methods ......................................................................................................... 18 References ..................................................................................................... 22 III. THE NECROTIC VENOM OF THE BROWN RECLUSE SPIDER INDUCES DYSREGULATED ENDOTHELIAL CELL-DEPENDENT NEUTROPHIL ACTIV ATION ................................................................................ 23 Introduction ................................................................................................. 23 Results ........................................................................................................... 25 Discussion ..................................................................................................... 34 References ..................................................................................................... 40 IV. ENDOTHELIAL CELL INFLAMMATORY RESPONSES TO TNFa: CERAMIDE-DEPENDENT AND INDEPENDENT MAP KINASE CASCADES .................................................................................................................. 44 Introduction ................................................................................................. 44 Results ........................................................................................................... 46 Discussion ..................................................................................................... 67 References ..................................................................................................... 75 V. ONCOSTATIN M INDUCES PROINFLAMMATORY AND MITOGENIC RESPONSES IN ENDOTHELIAL CELLS ..................................... 79 In.troduction ................................................................................................. 79 Results ........................................................................................................... 82 Discussion ..................................................................................................... 93 References ..................................................................................................... 97 VI. SIGNALS INDUCED BY CYTOKINES TNF AND OSM INTEGRATE AT INFLAMMATORY GENE PROMOTERS TO MODULATE TRANSCRIPTION ..................................................................................................... 99 In.troduction ................................................................................................. 99 Results .......................................................................................................... 101 Discussion .................................................................................................... 115 References .................................................................................................... 125 VII. TRANSFORMATION ALTERS TNF SIGNALING IN ENDOTHELIAL CELLS ............................................................................................ 127 In.troduction ................................................................................................ 127 Results .......................................................................................................... 129 Discussion .................................................................................................... 142 References .................................................................................................... 145 viii ACKNOWLEDGEMENTS I wish to thank Drs. Thomas McIntyre, my research advisor, Steve Prescott, and Guy Zimmerman for never ending patience, guidance, and support, throughout my stay at the lab. Importantly, Tom's, Guy's and Steve's encouragement and faith in my abilities helped me to overcome my feelings of scientific inadequacy that plagued me in my initial years in the lab. I am also grateful to my committee members for their helpful suggestions and exhortations to "wrap things up." Most of the work in this dissertation would not have been possible if not for expert technical assistance from Donelle Benson, Susan Cowley, Debbie Dykstra, Wen Hua Li, Aaron Ponstler, and Margaret Vogel. I am thankful to them and other colleagues at CVRTI for their help and friendship. I wish to thank my wonderful parents for their ceaseless and unconditional emotional and financial support during this most difficult period of my life. I am also indebted to Kimberly Dahms, Sudhakar Ihyer, Mark Martinez, and Kamala Patel for their friendship and humor. The work in this dissertation was funded by awards from Nora Eccles Treadwell Foundation and United States Public Health grant NIHLBI #P50 HL 50153. CHAPTER 1 INTRODUCTION Physiological basis of inflammation Inflammation is tissue response to injury that serves as protection against noxious agents. The cardinal features of inflammation are calor (heat), rubor (redness), dolor (swelling), and functio laesa (loss of function). These gross anatomical features are a result of coordinated intermolecular events occurring at the cellular level, initially between peripheral blood leukocytes and endothelial lining of the vascular wall. The vascular wall reacts by vasoconstriction within seconds of injury, followed by prolonged vasodilatation and margination of peripheral blood leukocytes over minutes to hours. Vasodilatation serves two purposes: it slows down blood flow and disrupts laminar flow to cause turbulence, thus allowing leukocytes, mainly polymorphonuclear neutrophil granulocyte (PMN), to come in contact with endothelium (1). EC, in the meanwhile, are activated to express on their surface P-selectin stored in Weibel-Palade bodies (2) and newly synthesized platelet activating factor (P AF) (3). This co expression of P-selectin, a tethering molecule, and platelet-activating factor, a signaling molecule, results in rolling of PMN on endothelium followed by tight binding (4). A similar scenario is seen when the offending stimulus is a gram negative bacterial injury (Fig. 1.1), in which the cell wall of bacteria containing lipopolysaccharide (LPS) activates endothelial cells to increase surface Unactivated PMN ~ CD11/CD18 ~ ,~ E-selectin ............ PMN Activation lit ( c :::> J EC. cC::;' ~ctivation "'-____ -' Unactivated EC 2 c Fig. 1.1 PMN adhesion to TNF or IL-1 stimulated endothelial cells requires coordinate expression of E-selectin and an activator like IL-8. TNF or IL-1 activates the expression of E-selectin to tether PMN to the endothelial cell. This tethering brings PMN in proximity to the coordinately expressed IL-8 bound to the EC surface. IL-8 activates PMN to increase avidity of CD11 /CD18 integrins to ICAM-1 and ICAM-2 on endothelial cells to mediate shear resistant tight binding. 3 expression of E-selectin, a tethering molecule, and IL-8, a signaling molecule. However, unlike P-selectin, the temporal sequence of events with E-selectin and IL-8 expression is much slower, occurring over hours rather than minutes, and requires gene induction (5). These events thus serve to localize leukocytes to a site of injury. The localization and activation of PMN serve one main purpose: transmigration of PMN across endothelial cells and home in on the offending stimulus by chemotaxis. PMN migration occurs via COII/COI8, an integrin complex expressed on the PMN surface, whose avidity of binding to endothelial surface ICAM-I and ICAM-2 or subendothelial matrix proteins is regulated by PMN activation induced by molecules such as IL-8 and PAF (6). Endothelial cells are important for localizing inflammatory responses and have the ability to mount an appropriate inflammatory response to varying types of stimuli. Despite the number of relevant stimuli, the evoked responses from the different receptors often is similar. For instance, TNF-a and IL-I bind unique receptors, but both evoke E-selectin and IL-8 expression in endothelial cells (7). This suggests that both IL-I and TNF-a send similar intracellular signals from their unique receptors to activate transcription, translation, or post­translational modification(s). To be able to pharmacologically modulate inflammatory responses, knowledge of both molecular interactions between leukocytes and endothelium and intracellular signaling inside leukocytes and endothelial cells is essential. The major focus of this study is the molecular mechanisms of endothelial signaling in response to inflammatory stimuli. A discussion of adhesion molecules, cytokines, and signaling relevant to this study is discussed below. 4 Adhesion molecules E-selectin. E-selectin is an adhesion molecule of the selectin family of transmembrane glycoproteins containing an extracellular ligand binding lectin domain, EGF receptor domain, complement repeats, and a single transmembrane domain and cytoplasmic tail (7). E-selectin interacts with L­selectin (8), PSGL-1 (9), ESL-1, and other unidentified ligand(s) on PMN surface. Quiescent endothelial cells do not express E-selectin, but upon activation by TNF, LPS or IL-1, surface expression of E-selectin is dramatically increased after 2 h, peaks between 4-8 h and subsides by 24 h (10,11). The transcriptional upregulation of E-selectin is achieved by activation of transcription factors NF-KB p50/p65, ATF-2, c-Jun and Hmg(I)Y to form a transcriptional initiation complex at the promoter region (12-14). NF-KB binding is essential for initiation of E-selectin expression and is augmented by ATF-2/ c-Jun heterodimer interactions (14). Regulation of one or more of these transcriptional factors by different agonists determine the degree of transcriptional activation of the E-selectin gene. ICAM-l. ICAM-1 is an adhesion molecule of the immunoglobulin family with five extracellular immunoglobulin-like domains, a single transmembrane domain, and a short cytoplasmic domain (6). It is expressed on the EC surface where it serves as a counter ligand for CD11/CD18 present on leukocytes, mediating shear-resistant binding prior to transmigration (15). Its expression is constitutive but is augmented by TNF, LPS and IL-1 via transcriptional upregulation following activation of promoter binding proteins NF-KB, p50/65, IRF-1 and C/EBP-a or C/EBP-fl (14). Depending on the agonist one or more of these factors are involved in augmenting transcriptional upregulation. 5 VCAM-l. VCAM-l, like ICAM-l, belongs to the immunoglobulin family of adhesion molecules, but unlike ICAM-l, it is not expressed on quiescent EC. Surface expression and transcriptional activation occur in response to TNF, IL-l, and LPS during which its main function is to recruit lymphocytes and monocytes via VLA4-VCAM-l interactions. The major transcription factors regulating VCAM-l expression are NF-lCB p50/p65 heterodimer and IRF-l. Each of these factors is cytokine activatable and results in early (~ h) but sustained expression over 24 h. This is in sharp contrast to E-selectin in which expression is downregulated within 24 h, suggesting regulation of NF-lCB by itself cannot account for these differences (14). Cytokines expressed by endothelial cells. Endothelial cells induce the expression of adhesion molecules and cytokines in a synergistic manner so that adhesion molecules localize and cytokines then activate these localized leukocytes. The major cytokines produced by endothelial cells in this context are IL-6 and IL-8, which activate leukocytes to undergo chemotaxis and degranulation. IL-6. IL-6, a pleotropic cytokine responsible for the acute phase response in bacterial infections, is produced by several cell types including endothelial cells (16). It is the prototype for the IL-6 family of helical cytokines, which share gp130 as a common subunit in a shared receptor complex (17). Agonists such as TNF-a, IL-1, and LPS induce IL-6 secretion from endothelial cells (18). IL-6 activates leukocytes (chiefly B-Iymphocytes), but it is also capable of activating PMNs to undergo chemotaxis and degranulation. Secretion occurs within 2h of exposure to stimulus, peaks by 6 6 h, and continues over 24 h by transcriptional upregulation induced by NF-KB and C/EBP~ (19,20). IL-8. IL-8 was isolated based on its ability to activate PMN and hence is also called neutrophil activating protein (21). It plays an important role in activation and chemotaxis of PMN by providing a juxtacrine activation signal for PMN localized by E-selectin on EC surface. In vivo it has been implicated as a mechanism for PMN activation in situations of oxidant stress, UV radiation, and rheumatoid arthritis (22). IL-8 is the prototypical member of the C-X-C chemokine family, containing two consecutive cysteines interrupted by an amino acid. Agonists such as TNF-a, IL-1, and LPS induce IL-8 secretion from endothelial cells after 2 h of stimulus with production at by 6-8 h and synthesis continuing over 24 h. Induction of mRNA occurs by modulation of NF-KB nuclear translocation and phosphorylation of C/EBP~, both of which bind to the promoter regions of the IL-8 gene (20,23-25). Though promoter region of the IL-8 gene binds the same proteins as IL-6, their transcriptional regulation is substantially different, a topic addressed in this dissertation. Signaling Changes in endothelial cell behavior induced by proinflammatory stimuli occur by transduction of signals from the outside to the inside of the cell. These signals consist of different cascades that bring about rapid and transient responses, resulting in amplification of signal without loss of specificity. Only signaling mechanisms that are relevant to this course of study are discussed. MAP kinase signaling. Mitogen activated protein (MAP) kinases were initially identified as kinases that underwent activation in response to growth 7 factors and oncogenic signals. The MAP kinase family members are characterized by phosphorylation on both tyro sines and threonines by an upstream dual specificity threonine and tyrosine kinase. They are broadly divided into distinct, but perhaps, overlapping cascades (Fig. 1.2) (26).The classical Ras/ERK pathway occurs by activation of Ras upon GTP binding, which triggers membrane association of Raf leading to its activation by unidentified mechanisms. Raf phosphorylates the dual specificity kinase MEK, on a serine, which activates it and causes it to phosphorylate ERKs (Extracellular Signal Regulated Kinase) on threonine and tyrosine residues. The resulting activation of ERKs induces phosphorylation of transcription factors and downstream kinases leading to changes in gene expression of target genes (26,31). Though many of the downstream targets are yet to be identified, C/EBP-~ has been identified to modulate transcription of IL-6 and IL-8 upon phosphorylation downstream of ERK 1 and 2. A newly described pathway most often induced by cellular stress results in activation of MEKK independent of Ras and Raf-l, which activates dual specificity kinases MKK3 and MKK4. MKK3 activates the MAP kinase p38, and MKK4 activates both the MAP kinases JNK-l (SAPK-l) and p38( SAPK-2) (Fig. 1.2). JNK-l and p38phosphorylate c-Jun and ATF-2 to induce transcription (29,32,33). Both these transcription factors bind to the promoter region of E-selectin gene where they are involved in transcription initiation in conjunction with the transcription factors of the NF-KB family. Lipid signaling. Generation of lipid second messengers is a powerful tool utilized by cells to transduce signals as the precursors are stored away from the enzymes in a high density state. Signaling is achieved by activation of signal-activated phospholipases that cleave membrane phospholipids to produce products that are second messengers by themselves or are substrates 8 c: Stimulus :> ~ ~ ~ MAP KKK Raf ? MEKK l l t MAPKK MEK MKK3 MKK4 (SEK, JNKK) E~K pta f MAPK , JNK (SAPK) t , transcription factor activation , , , Gene Induction Fig 1.2 Mammalian MAP kinase cascades: The outline of MAP kinases in mammalian cells is a composite (27-30) and is meant to convey a current understanding of the cascades, their relationship to one another, and alternative nomenclature. The growth-factor-induced pathway results in activation of ERI< whereas the p38 and JNK-l MAP kinases are activated by cellular stress. 9 for down stream second messenger producing enzymes. This mechanism has mainly been described in the context of phospholipases, and the major focus of this study is the role of signal-activated sphingomyelinase in the hydrolysis of membrane sphingomyelin in response to different signals like TNF, 1,25 dihydroxy cholecalciferol (Vitamin D3 or 1,25 OH D3), Nerve Growth Factor, and Fas ligand (35,36). Ceramide thus produced is capable of activating NF-lCB translocation to the nucleus (37), cell differentiation, cell cycle arrest, and apoptosis depending on the cell type and conditions (Fig. 1.3). Downstream ceramide targets are not well characterized, but a few candidates have been identified: Ceramide-activated protein kinase (CAPK) is associated with activation of the Raf/ERK pathway at the level of Raf-1 (38), a ceramide-activated protein phosphatase (CAPP) that belongs to the protein phosphatase of the IIA class dephosphorylates unidentified proteins that are phosphorylated serine/threonine (39,40), and PKC ~ that catalyzes IlCB phosphorylation in vitro which would lead to NF-lCB translocation in vivo (41) A confounding factor in the identification of ceramide targets is that activatable sphingomyelinase activity may be topology specific in some cell types: stimulated sphingomyelin hydrolysis by plasma membrane associated sphingomyelinase results in ceramide generation in the cell membrane leading to MAPK activation, whereas activation of a lysosomal sphingomyelinase by diacylglycerol results in intracellular ceramide generation, which in turn leads to NF-lCB translocation into the nucleus (34). The two different ceramide pools thus produced have been reported to be separate, suggesting a further level of complexity (Fig. 1.4). The initiation of ceramide stimulus and its effects on endothelial cells will be discussed in detail in Chapter 3. o H Membrane Sphingomyelin TNF-R Ceramide -~ ....... Phosphate ? ~ H + Choline Second Messenger?? HO H 10 o + Choline-phosphate Fig 1.3 Sphingomyelin hydrolysis pathways. Cell membrane sphingomyelin can be hydrolyzed by a sphingomyelinase C (SMC) to generate ceramide or a sphingomyelinase D (SMD) to generate ceramide phosphate. TNF receptor (TNF-R) couples to a sphingomyelinase C to transduce signal. SMD is currently unknown in mammalian biology. 11 :~ 1 PLA2 cera, / MAP kinase DAG TNF-R /\ PKC Acidic SMase • Ceramide I Arachidonic Acid NF-x:B Fig 1.4 Proposed topology specific ceramide effects (34). Two ceramide signals may exist in some cells. Plasma membrane generated ceramide activates MAP kinase activity that leads to phospholipase A2 (PLA2) activation. Internal ceramide activates a phosphatidylcholine specific phospholipase C (PC-PLC) that produces diacylglycerol (DAG). This leads to activation of protein kinase C (PKC) and acidic sphingomyelinase to cause NF-x:B translocation. 12 NF-KB signaling. One of the major mechanisms of pro inflammatory gene induction in endothelial cells involves the nuclear translocation of NF­KB (Fig. 1.5), a family of transcription factors constitutively expressed and localized to the cytoplasm by IKB in quiescent endothelial cells (14,42). The IKB/NF-KB complex masks the nuclear localization sequence of NF-KB to prevent nuclear translocation. Stimulation with TNF, LPS, or IL-l leads to rapid phosphorylation of IKB and ubiquitin-dependent degradation by the proteosome complex in the cytoplasm. The unmasking of NF-KB nuclear localizing sequences after IKB degradation allows nuclear translocation of NF­KB (43). Here it binds promoter regions of E-selectin ICAM-l, VCAM-l, IL-6, and IL-8 (14). Different members of the NF-KB family have been identified: p50 (NF-KB1) is synthesized as a pl05 protein that is bound to p65 (ReI A) (44). Stimulation with TNF or IL-l causes degradation of the C-terminus to produce p50/p65 heterodimers that translocate to the nucleus to allow gene transcription. Similarly p52 (NF-KB2) is synthesized as pl00 protein bound to p65 whose C-terminus is degraded upon activation leading to nuclear translocation of p52/p65 heterodimer. IKB proteins also consist of a family of proteins characterized by the presence of ankyrin repeats involved in regulation of localization of NF-KB (44). The predominant forms are IKB-a, IKBy (the C-terminus of pl05) and the C-terminus of pl00. References 1. Marchesi, V. T. (1985) in Anderson's Pathology (Kissane, J. M., ed) pp. 22- 60, C.V. Mosby, St. Louis 2. Lorant, D. E., Topham, M. K., Whatley, R. E., McEver, R. P., McIntyre, T. M., Prescott, S. M., and Zimmerman, G. A. (1993) J. Clin. Invest. 92, 559-570 Stimulus p105! ~65 r;:I5 <Ja~ /hcBa • protein klnase(s) • p105! 65 ~ P~O/ 65 Cl proteosome mediated p105 C-terminus and hcBa degradation hcBa CJi5 • ----------------Ci ai5 NF·KB binding she 13 Cytoplasm Nucleus Fig 1.5 Pathways utilized for NF-kB translocation. Certain stimuli like TNF, IL-1 and LPS, activate a protein kinase(s) activity that phosphorylates the C­terminus of p105 or IKB-a and tags them for proteosome mediated degradation. Degradation results in unmasking of the nuclear localizing sequence of p65, causing translocation of SO/p65 heterodimers. Figure modified from (14). 3. Zimmerman, G. A., Prescott, S. M., and McIntyre, T. M. (1992) in Inflammation: Basic Principles and Clinical Correlates (Gallin, J. 1. and Goldstein,1. M., eds) pp. 149-176, Raven Press, New York 14 4. Lorant, D. E., Patel, K. D., McIntyre, T. M., McEver, R. P., Prescott, S. M., and Zimmerman, G. A. (1991) J. Cell Bioi 115, 223-234 5. Pober, J. S. and Cotran, R. S. (1995) Physiol. Rev 70, 427-451 6. Diamond, M. S., Staunton, D. E., Marlin, S. D., and Springer, T. A. (1991) Cell.65, 961-971 7. Bevilacqua, M. P., Corless, C., and Lo, S. K. (1991) in Cellular and Molecular Mechanisms in Inflammation: Vascular Adhesion Molecules (Cochrane, C. G. and Gimbrone, M. A., eds) pp. 1-13, Academic Press, Inc., San Diego 8. Jutila, M. A., Kishimoto, T. K., and Finken, M. (1991) Cell. Immunol. 132, 201-214 9. Polley, M. J., Phillips, M. L., Wayner, E., Nudelman, E., Singhal, A. K., Hakomori, S. -I., and Paulson, J. C. (1991) Proc. Natl. Acad. Sci. USA. 88, 6224- 6228 10. Larsen, G. R., Sako, D., Ahern, T. J., Shaffer, M., Erban, J., Sajer, S. A., Gibson, R. M., Wagner, D. D., Furie, B. C., and Furie, B. (1992) J. BioI. Chern. 267, 11104-11110 11. Montgomery, K. F., Osborn, L., Hession, C., Tizard, R., Goff, D., Vassallo, C., Tarr, P. I., Bomsztyk, K., Lobb, R., Harlan, J. M., and Pohlman, T. H. (1991) Proc. Natl. Acad. Sci. USA. 88, 6523-6527 12. van Huijsduijnen, R. H., Whelan, J., Pescini, R., Becker-Andre, M., Schenk, A. -M., and DeLamarter, J. F. (1992) J. BioI. Chern. 267, 22385-22391 13. Kaszubska, W., van Huijsduijnen, R. H., Ghersa, P., DeRaemy-Schenk, A. -M., Chen, B. P. C., Hai, T., DeLamarter, J. F., and Whelan, J. (1993) Mol. Cell. BioI. 13, 7180-7190 14. Collins, T., Read, M. A., Neish, A. S., Whitley, M. Z., Thanos, D., and Maniatis, T. (1995) FASEB. J 9, 899-909 15. Smith, C. W., Marlin, S. D., Rothlein, R. R., Toman, C. J., and Anderson, D. C. (1988) Blood 72, 1-1531(Abstract) 16. Kopf, M., Baumann, H., Freer, G., Freudenberg, M., Lamers, M., Kishimoto, T., Zinkernagel, R., Bluethmann, H., and Kohler, G. (1994) Nature 368,339-342 15 17. Hirano, T., Matsuda, T., and Nakajima, K. (1994) Stem Cells 12, 262-277 18. Isshiki, H., Akira, S., Tanabe, 0., Nakajima, T., Shimamoto, T., Hirano, T., and Kishimoto, T. (1990) Mol. Cell. BioI. 10, 2757-2764 19. Libermann, T. A. and Baltimore, D. (1990) Mol. Cell. BioI. 10, 2327-2334 20. Matsusaka, T., Fujikawa, K., Nishio, Y., Mukaida, N., Matsushima, K., Kishimoto, T., and Akira, S. (1993) Proc. Nati. Acad. Sci. USA. 90, 10193-10197 21. Huber, A. R., Kunkel, S. L., Todd, R. F.,III, and Weiss, S. J. (1991) Science 254,99-102 22. Terkeltaub, R., Banka, C. L., Solan, J., Santoro, D., Brand, K., and Curtiss, L. K. (1994) Arterioscler. Thromb. 14, 47-53 23. Mahe, Y., Mukaida, N., Kouji, K., Akiyama, M., Ikeda, N., Matsushima, K., and Murakami, S. (1991) J. BioI. Chem. 266, 13759-13763 24. Stein, B. and Baldwin, A. S.,Jr. (1993) Mol. Cell. BioI. 13, 7191-7198 25. Mukaida, N., Mahe, Y., and Matsushima, K. (1990) J. BioI. Chern. 265, 21128-21133 26. Cano, E. and Mahadevan, L. C. (1995) Trends. Biochem. Sci. 20, 117-123 27. Yan, M., Dai, T., Deak, J. C., Kyrlakis, J. M., Zon, L. I., Woodgett, J. R., and Templeton, D. J. (1994) Nature 372, 798-800 28. Lin, A., Minden, A., Martinetto, H., Claret, F. -X., Lange-Carter, C., Mercurio, F., Johnson, G. L., and Karin, M. (1995) Science 268, 286-290 29. Derijard, B., Raingeaud, J., Barrett, T., Wu, I. -H., Han, J., Ulevitch, R. J., and Davis, R. J. (1995) Science 267, 682-684 30. Marshall, C. J. (1995) Cell 80, 179-185 31. Lange-Carter, C. A., Pleiman, C. M., Gardner, A. M., Blumer, K. J., and Johnson, G. L. (1993) Science 260, 315-319 32. Kyriakis, J. M., Banerjee, P., Nikolakaki, E., Dai, T., Rubie, E. A., Ahmad, M. F., Avruch, J., and Woodgett, J. R. (1994) Nature 369, 156-160 33. Sanchez, I., Hughes, R. T., Mayer, B. J., Yee, K., Woodgett, J. R., Avruch, J., Kyriakis, J. M., and Zon, L. I. (1994) Nature 372, 794-798 34. Wiegmann, K., Schutze, S., Machleidt, T., Witte, D., and Kronke, M. (1994) Cell 78, 1005-1015 16 35. Pushkareva, M., Obeid, L. M., and Hannun, Y. A. (1995) Immunol. Today. 16,294-297 36. Hannun, Y. A. (1994) ,. BioI. Chern. 269, 3125-3128 37. Belka, C., Wiegmann, K., Adam, D., Holland, R., Neuloh, M., Herrmann, F., Kronke, M., and Brach, M. A. (1995) EMBO. '.14, 1156-1165 38. Liu, J., Mathias,S., Yang, Z., and Kolesnick, R. N. (1994) ,. BioI. Chern. 269, 3047-3052 39. Dobrowsky, R. T. and Hannun, Y. A. (1992) ,. BioI. Chern. 267, 5048-5051 40. Dobrowsky, R. T., Kamibayashi, C., Mumby, M. C., and Hannun, Y. A. (1993) J. BioI. Chern. 268, 15523-15530 41. Lozano, J., Berra, E., Municio, M. M., Diaz-Meco, M. T., Dominguez, 1., Sanz, L., and Moscat, J. (1994) ,. BioI. Chern. 269, 19200-19202 42. Read, M. A., Whitley, M. Z., Williams, A. J., and Collins, T. (1994) ,. Exp. Med.179,503-512 43. Reddy, S. A. G., Chaturvedi, M. M., Darnay, B. G., Chan, H., Higuchi, M., and Aggarwal, B. B. (1994) ,. BioI. Chern. 269, 25369-25372 44. Blank, V., Kourilsky, P., and Israel, A. (1992) Trends. Biochern. Sci 17, 135- 140 CHAPTER 2 MATERIALS AND METHODS Materials Venom, collected from the fangs of Loxosceles deserta by electrical stimulation of the spiders (to prevent contamination with gastric contents), was obtained from Spider Pharm (Feasterville, P A [phone]). Venom was stored as small aliquots at -800 ; it was thawed, diluted 100-fold with HBSS containing 0.5% human serum albumin (HBSS/ A) and centrifuged in a microfuge to remove precipitated material. A fresh lot was prepared weekly as thawed venom loses activity over a period of several days and is inactivated by a single round of freeze/ thaw after this dilution. Recombinant human TNF was provided by Genentech (South San Francisco, CA). Staphylococcus aureus sphingomyelinase C was obtained from Sigma Immunochemicals (St. Louis, MO). HBSS and M199 were obtained from Whittaker Bioproducts (Walkersville, MD), and human serum albumin was from Miles Laboratories (Elkhart, IN). N-Octanoyl sphingosine were a gift from Robert Bell (Duke University, Durham NC) and Avanti Polar Lipids respectively. Mouse antihuman E-selectin antibody 3E6 was a gift from Rodger McEver and blocking P6E2 anti-E-selectin antibody from James Paulson (Cytel). Oncostatin M, Polyclonal rabbit antihuman IL-8 and recombinant human IL-8 were obtained from R&D Systems (Minneapolis, MN), and mouse antihuman GM-CSF, polyclonal rabbit antihuman GM-CSF, and recombinant human GM-CSF were from Genzyme (Cambridge, MA). Endogen 18 (Boston, MA) supplied polyclonal rabbit antihuman IL-6 and Biosource International (Camarillo, CA) supplied recombinant human IL-6. Mouse monoclonal anti-VCAM-1 was obtained from Biodesign International. (Kennebunkport, ME). Recombinant c-Jun andoligonucleotide for NF-lCB were purchased as a part of the gel shift assay core system from Promega Corporation (Madison WI). ECL western blotting reagents, 32p-ATP, DAG assay kit, and MAP kinase assay kit were purchased from Amersham Corporation (Arlington Heights, IL) recombinant ATF-2, recombinant MEK-1, and all kinase antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). All secondary antibodies were obtained from Biosource international (Camarillo, CA). Other reagents were from Sigma chemical company (St. Louis, MO). Methods Cell culture adhesion assays and imaging. Human umbilical vein endothelial cells were cultured in 24 mm or 35 mm multiwell plates (Costar Data Packaging Corp., Cambridge MA) as described (1). Only monolayers of primary cultures that were tightly confluent were used for these studies. Neutrophils were isolated from human blood and labeled with lllIn-oxine as described (1). Endothelial cells were treated with the stated agonist in HBSSI A for the specified period of time, washed with HBSS, and the percentage of lllIn-neutrophils that tightly adhered in 5 min was quantified as described (2). For short-chain ceramide treatment of endothelial cells, the desired quantity of 10-2M N­octanoylsphingosine or N-hexanoylsphingosine or N-acetylsphingosine solution in ethanol was added to a test tube with sterile HBSSI A and rocked at 370C for 1 h prior to addition to cells. An ethanol vehicle control was performed whenever ceramide was used. When the effect of anti-E-selectin antibody on PMN­endothelial cell interactions was examined, the appropriate cell was preincubated 19 I with 10 /lg/ml of the mAb for 30 min and the adhesion assay performed in the presence of this concentration of antibody. A positive control for antibody effectiveness was included in each experiment. All adhesion values are reported as the mean and range of duplicate values and are representative of at least two experiments. Ceramide adhesion assay performed by capturing the microscopic image with a video camera onto video tape. A framegrabber with NIH image 1.47 was used to obtain a single image. Flow cytometry , ELISA, and western blotting of endothelial cells. Surface expression of E-selectin or ICAM-1 was quantitated by removing treated or control endothelial cells by swirling them for 3 min at room temperature on an orbital shaker at 200 rpm in the presence of Trypsin-EDTA (GIBCD BRL). A few persistently adherent cells were removed by gently repipeting the contents of the culture dish. The cells were washed twice in 50% goat serum and stained with anti-E-selectin (P6E2) or ICAM-1 (18E3D) or VCAM-1(Biodesign) monoclonal antibody in this buffer. FITC-conjugated goat anti-mouse (Sigma) in 50% goat serum was used as the secondary antibody followed by fixation in 0.5% formaldehyde. Viability of recovered unfixed cells was examined in initial experiments in parallel using propidium iodide stain; experience gained from previous experiments was used to gate FITC-Iabeled cells in future experiments. Material for western blots was collected by solubilizing monolayers with boiling Laemmli sample buffer without J3-mercaptoethanol. Proteins were electorphoretically separated under nonreducing conditions on 7.5% SDS polyacrylamide gels and transferred to a PVDF membrane (Immobilon P, Millipore Corporation, Bedford, MA) (3). The membrane was blocked with 5% nonfat milk in Tris buffered saline with 0.05% Tween-20 (TBST) and probed with 1Jlg/ml anti-E-selectin monoclonal antibody 3E6 (gift from Rodger McEver). Horseradish peroxidase-conjugated polyclonal goat antimouse (1:5000) in the 20 presence of 100 J.1g/ ml nonimmune rabbit IgG in TBST was used in secondary antibody incubation. Staining was detected with ECL. NIH Image 1.47 was used to estimate the density of the scanned bands. IL-6 and IL-8 were quantitated from endothelial cell supernatants by sandwich ELISA with polyclonal rabbit antihuman antibody; detection employed biotinylated rabbit antihuman IL-8 or IL-6 and avidin-conjugated horseradish peroxidase. Preparation of nuclear extracts. Endothelial cells were treated for the stated times and washed twice with ice cold phosphate buffered saline (PBS). The cells were removed by scraping with a cell scraper and pelleted by centrifuging at 300 g for 5 min at 40 C. The cells were then resuspended in 1 ml ofbufferA(10mMHEPESpH7.9, 10mMKCI, lmMEDTA, lmMEGTA, 1 mM PMSF, 1 mM DTT, 10 J.1g/ml aprotinin, 100 J.1M leupeptin, and 0.5% NP-40) for 10 min on ice, vortexing every 2 min. Nuclei were collected by centrifugation at 500 g for 10 min at40 C. The nuclear pellets were washed 3 times with 500 ml buffer A without NP-40. To the nuclear pellet 50 ml of buffer B ( 10 mM HEPES pH 7.9, 420 mM NaCI, 25% glycerol,S mM MgCI2, 0.1 mM EDTA, 0.1 mM EGTA, 10 J.1g/ml aprotinin, 100 J.1M leupeptin, 1 mM PMSF, and 1 mM DTT) was added, sonicated for 5 sec in a Branson sonicator and incubated for 30 min on ice with vortexing every 5 min. Nuclear debris was removed by centrifugation at 13,000 g for 10 min at 40 C. The supernatant was collected and analyzed immediately as described below. Electrophoretic mobility shift assays. Equal amounts of nuclear extracts -3 J.1g of protein (determined by Bradford protein assay) were incubated with -30,000 cpm (37.5 pMoles) of 32p labeled NF-1CB specific probe (Promega). Reactions were performed in a 30 J.11 volume containing 7.5 J.11 nuclear extract, 6 J.11 5X gel shift binding buffer (20 mM tris pH 7.9, 5 mM MgCI2, 0.5 mM DTT, 0.5 mM EDTA and 200/0 glycerol), 1.5 J.1g Poly dI-dC and 3.75pMol Oct-l 21 oligonucleotide as non-specific competitor DNA. The reaction was incubated at room temperature for 15 min and separated on a 4% native polyacrylamide gel in O.5X TBE buffer. The gel was dried and subjected to autoradiography. Inclusion of 100-fold excess of the specific unlabeled oligonucleotide abrogated formation of DNA-probe complexes on autoradiography (4). DAG Kinase assay. Lipids from 2 X 106 endothelial cells were extracted by the Bligh and Dyer technique (5) at the stated times after treatment and 25-50 % of the lipid was subjected to diacylglycerol kinase assay as described by the manufacturer with a few modifications: 1) Only glass tubes were used. 2) C-8 ceramide was included as standard in addition to diacylglycerol. 3) Reaction time was extended to 3 h. 4) Completed reaction mixture was separated on a HP-TLC and autoradiography was performed to locate the bands. 32p labeled ceramide phosphate and phosphatidic acid bands were scraped and counted on a liquid scintillation counter. MAP and Kinase Raf-l assays. Endothelial cells cultured in costar 6-well plates were serum starved for 24-36 h in serum free M199 with 0.5% human serum albumin. After treatment with agonist for the stated times, cells were washed once in ice cold PBS and placed on a bed of ice. Cells were scraped, pelleted, and resuspended in 150 ~ of kinase buffer (20 mM HEPES pH 7.4, 150 mM NaCI, 20 mM MgC12, 25 mM ~-glycerophosphate, 7.5 mg/ml human serum albumin, 2 mM OTT, 2 mM sodium orthovanadate, 10 J.1g/ml aprotinin, 100 JlM leupeptin, 1 mM PMSF, 25 J.1M ATP) by passage through a 25G needle 10 times. Insoluble debris was cleared by centrifugation at 12,000 g for 5 min. For Jun kinase assays, 100 ng of recombinant human c-jun with -2.5 nCi of ATP in 15 ~ of kinase buffer was added to 15 J.11 of cell lysate and incubated for 30 min at 37°C (6). For Raf-l kinase assay, 20 ~ of Protein G agarose beads (Sigma) previously incubated for 2 h at 4°C with 3 J.1g of Raf-l antibody and extensively washed was 22 added to 75 J.lI of celllysates for 1h at 4°C (7). Beads were subsequently pelleted by centrifugation and washed 3 X 500 J.lI in kinase buffer and incubated at 370C for 1 h with 30 J.lI of kinase buffer containing 1 J.lg recombinant human MEK-1 and 2.5 nCi of 32p-ATP. The kinase assay reactions were stopped by addition of 10 j..Ll4X Laemmli sample buffer and boiling for 3 min. This was analyzed following SDS-P AGE by autoradiography. ERK kinase assay was performed using Amersham MAP kinase assay kit according to manufacturers instructions. Kinase and ATF-2 western blots. Celllysates and pellets used for kinase assays were homogenized in Laemmli sample buffer and separated on 9% SDS­PAGE and transferred to Immobilon-P membrane (3). The membranes were probed for ERK, p38, and JNK-1 according to manufacterers instructions accompanying the primary antibody. Blots were developed using ECL. References 1. Zimmerman, G. A., McIntyre, T. M., and Prescott, S. M. (1985) J. Clin. Invest. 76, 2235-2246 2. Zimmerman, G. A., McIntyre, T. M., Mehra, M., and Prescott, S. M. (1990) J. Cell. BioI. 110,529-540 3. Laemmli, U. K. (1970) Nature 227, 680-685 4. Read, M. A., Whitley, M. Z., Williams, A. J., and Collins, T. (1994) J. Exp. Med. 179, 503-512 5. Bligh, E. G. and Dyer, W. J. (1959) Can. J. Biochem. Physiol. 37,911-917 6. Gupta, S., Campbell, D., Derijard, B., and Davis, R. J. (1995) Science 267, 389-393 7. Minden, A., Lin, A., Lange-Carter, C., Derijard, B., Davs, R. J., Johnson, G. L., Karin, M., and McMahon, M. (1994) Science 266, 1719-1723 CHAPTER 3 THE NECROTIC VENOM OF THE BROWN RECLUSE SPIDER INDUCES DYSREGULATED ENDOTHELIAL CELL-DEPENDENT NEUTROPHIL ACTIVATION In trod uction Envenomation by the brown recluse spider (Loxosceles reclusa ), its desert counterpart (Loxosceles deserta) (1,2), or the South American Loxosceles Laeta (1) can result in an impressive dermonecrotic lesion. This progresses from an acute local inflammatory reaction to a black eschar, in which sloughing of the necrotic tissue leaves a draining, well-demarcated ulcer (3). These lesions are remarkable considering that these spiders emit only a few tenths of a microliter of venom (4). The mechanism by which the venom causes necrotic lesions is currently unknown. The polymorphonuclear leukocyte (PMN, neutrophil) might be expected to be the central target of the venom because depletion of PMN with a single dose of nitrogen mustard delays hemorrhage, edema, and necrosis until the pool of circulating neutrophils is reestablished (5). This key observation shows that PMN are the likely proximal cause of the tissue destruction associated with the intense inflammatory reaction following envenomation. However, human neutrophils are not themselves activated by the venom ex vivo and, in fact, are actually inhibited by it (6,7). This apparent paradox may be resolved by the observation (8) that the first ultrastructural change after envenomation 24 is selective damage to vascular endothelium, whereas at the level of light microscopy it is adhesion of neutrophils to the capillary wall (9). This suggests that an essential component of this inflammatory reaction may be activation of vascular endothelium, with subsequent sequestration and activation of passing neutrophils by the perturbed endothelial cells. The initial step in leukocyte emigration from the blood stream is their interaction with the endothelium. This may occur either by direct activation of neutrophils, which functionally upregulates their CDll/CD18 adhesive glycoprotein complex, or by activation of endothelial cells to express neutrophil tethering and activating molecules (10). Adhesion by the activated CDll/CD18 integrins is sensitive to the shear (11) of flowing blood and in the absence of stasis is not likely to be the initial event in leukocyte interaction with the vascular wall. In contrast, endothelial cell-dependent adhesion, by virtue of the adhesion molecules presented by activated endothelial cells, is resistant to this. There is a general paradigm by which activated endothelial cells bind neutrophils: two inflammatory processes use different molecules to accomplish similar functions (10). Agonists like throrrlbin and histamine within minutes induce the translocation of P-selectin from intracellular granules, Weibel-Palade bodies, to the apical endothelial cell surface. Surface P-selectin acts to tether neutrophils to the endothelial cells by binding to a sialylated counter receptor on neutrophils. Simultaneously, endothelial cells synthesize platelet-activating factor (P AF) and translocate it to their plasma membrane, where it can activate the tethered neutrophils through a specific receptor. This juxtacrine stimulation ensures that only adherent neutrophils are stimulated to extravasate. Similarly, endothelial cells exposed to TNF, IL­l, or endotoxin begin to synthesize and then express a related tethering molecule, E-selectin (10,12). This process occurs over a period of hours as 25 neither E-selectin nor its mRNA is detectable in unactivated endothelial cells. Tethered PMN are activated by cytokine-induced endothelial cells, and, although the nature of the neutrophil stimulatory molecule(s) in this setting is not completely defined, cytokine-activated endothelial cells also begin to synthesize and release the PMN agonist IL-B. Previously we tested the hypotheses that Loxosceles venom is an endothelial cell agonist and that these activated endothelial cells in turn activate PMN. We exposed primary cultures of human umbilical vein endothelial cells to minute amounts of Loxosceles venom and found that by 4 h treated monolayers bound and activated human neutrophils. However, the venom differentially induced endothelial cell inflammatory events. PMN adhered only to the junctions of venom-activated endothelial cells, and they did not alter their morphology to a polarized form nor did they migrate beneath the monolayer. Despite this lack of motile responses, adherent PMN released contents from both primary and secondary granules. In the corresponding paper she hypothesized that this dysregulated inflammatory response may account for the tissue destruction associated with envenomation by Loxosceles spiders. My investigations revealed that, venom treated monolayers supported PMN adhesion via E-selectin despite its low surface expression and that this correlated to low ICAM-l and VCAM-l expression. The main mechanism for poor agonist activity is suboptimal NF- 1(B translocation induced by BRSV. Results Loxosceles venom activates endothelial cells. to bind PMN via E­selectin. We exposed primary cultures of endothelial cells to varying amounts of Loxosceles venom for 4 h, washed them, and then incubated 26 freshly isolated, unactivated PMN with these monolayers. We found that PMN bound to venom-activated endothelial cells, and that this supported PMN adhesion of significantly inhibited by P6E2, an E-selectin specific antibody (Fig. 3.1). This monoclonal antibody has been epitoped mapped (13) to a small region the lectin-binding domain responsible for carbohydrate recognition, indicating that the same critical region of E-selectin is involved in PMN tethering following its induction by either TNF or spider venom. On quantitating E-selectin immunoreactive material following electrophoretic separation and probing of the blotted proteins with an anti-E-selectin mAb, we found that venom-exposed cells expressed E-selectin, and that it co­migrated with E-selectin expressed by cytokine-activated cells, but Loxosceles venom was only a weak agonist for E-selection expression: quantitation of the developed blots showed that venom-exposure induced less than 10 % of the E-selectin expressed by TNF-treated endothelial cells (Fig. 3.2). This ratio of immunoreactive protein was maintained to at least 8 h of exposure (not shown), indicating that this difference was not merely due to a slower rate of induction. Finally, we quantitated E-selectin surface expression by FACS analysis. We found (Fig. 3.3) that E-selectin expression was significantly above the nonspecific staining of control cells, but was well below the levels expressed by cytokine-stimulated cells. Together these approaches have demonstrated that Loxosceles venom induces E-selectin expression, but is a far less potent agonist than TNF for this process. Venom-activated endothelial cells differentially express neutrophil agonists. Endothelial cell-dependent PMN adhesion results from both from expression of a tethering molecule and a signaling molecule to induce functional responses from the bound PMN (10,14). We considered the possibility that the lack of altered PMN morphology and migration below 27 .. Control f2) Anti-E-selectin 60 50 40 %PMN 30 Binding 20 10 0 Control TNF BRSV Fig 3.1 E-selectin monoclonal antibodies block PMN adhesion to Loxosceles venom-activated monolayers. Confluent monolayers were treated for 4 h with buffer, 250 Vlml TNF, or O.If..Ll BRSV Iml, the incubation medium was removed and the monolayer washed with buffer, and then either HBSSI A or 10 f..Lg/ml of P6E2 anti-E-selectin mAb were added. After 30 min at 37° the antibody was removed, and either neutrophils alone (control) or neutrophils (containing 10 f..Lg/ml of the antibody used to pretreat the monolayer to maintain its concentration during the coincubation) were then added. Neutrophil adhesion after the 5 min coincubation was determined. Results from control incubations with an irrelevant antibody were not different from buffer-treated controls. Control 1 1 28 TNF BRSV 0.2 0.1 1 Fig.3.2 E-selectin is expressed in Loxosceles venom-activated HUVEC. HUVEC were treated for 4 hat 3J70C with HBSS/ A alone or with either 250U/ml TNF, or O.lfll BRSV /ml. After the incubation the monolayers were washed and E-selectin expression was determined by immunoblotting. 29 E-Selectin --Control -TNF ---BRSV Fig 3.3 FACS analysis of endothelial cell E-selectin. Endothelial cells were treated with buffer, TNF, or 0.3 J.LI/ml Loxosceles venom for 4 h, removed from their substrate, stained with E-selectin monoclonal antibody (BBA7) or control antibody. Staining of viable cells was determined as described in "Materials and Methods." These experiments are typical of two other analyses. 30 venom-treated endothelial cell mono layers was due to the expression of a tethering molecule in the absence of an appropriate neutrophil agonist. This seemed possible given the weak expression of E-selectin described above. Accordingly, we determined if venom-activated endothelial cells synthesized and released IL-8, an early response gene product which is a potent mediator of neutrophil activation and extravasation (15-17). We collected supernatants of monolayers treated with TNF or Loxosceles venom and quantified IL-8 by ELISA. We found that, in contrast to the weak induction of E-selectin, venom-exposed mono layers were significant producers of this cytokine: IL-8 production was stimulated 10-fold by the venom, compared to a 22-fold induction in response to TNF (Fig. 3.4). Moreover, we found a second cytokine, GM-CSF, was induced by venom exposure, and unexpectedly that its accumulation by 4 h was far in excess of that expressed by TNF-exposed monolayers: Loxosceles venom induced a 6.4-fold increase in GM-CSF release compared to a 1.3-fold increase induced by TNF. This result was unexpected as stimulation of GM-CSF release is slow; elaboration in response to TNF is detectable only by 8 h of exposure (18). Elaboration of this cytokine would also have functional consequences for adherent PMN since it is a direct PMN agonist (19-21), as well as a potent priming agent (22). In contrast to the elaboration of IL-8 and GM-CSF, venom stimulation of endothelial cell mono layers did not induce the synthesis and release of IL-6 (Fig. 3.4). Since monolayers treated with TNF did show a major induction of IL-6 synthesis, this response clearly shows that venom-activated monolayers differentially express inflammatory gene products compared to this well-established endothelial cell agonist. Expression of two cytokines that stimulate and/or prime PMN function also falsifies the postulate that venom-activated monolayers retain PMN on their apical surface due to an inability to 1 1 IL-8 Control TNF BRSV ng/ml 0 0.3 0.2 ng/ml 0.1 31 IL-6 Control TNF BRSV GM-CSF Control TNF BRSV Figure 3.4 Loxosceles venom-activated endothelial cells release PMN agonists IL-8 and GM-CSF but not IL-6. HUVEC were treated for 4 h with HBSS/ A, TNF, or 0.3 ul/ml of Loxosceles venom/ml. After the incubation the buffer was removed, and the supernatants were assayed for IL-8, IL-6, or GM-CSF by sandwich ELISA. The endothelial cell monolayers were then used in the PMN adhesion assay described in Fig. 3.1 as a positive control: the increased amount of venom in these experiments produced the same level of PMN adhesion, determined in parallel, to venom- and TNF-activated monolayers. These results are representative of two other identical experiments. 32 appropriately activate them. Venom induced monolayers fail to activate ICAM-l and VCAM-l expression. ICAM-l expression is constitutive but augmentable on the EC surface. Augmentation by TNF coordinated with E-selectin surface expression allows for efficient presentation of a CDll/CD18 counter ligand to promote shear resistant binding and transmigration. Since spider-venom­treated endothelial cells do not support PMN transmigration, we determined whether failure to augment ICAM-l could partially account for this. We determined cell surface expression of ICAM-l after exposure of ECs for 4 h to spider venom, and found it was a poor agonist for ICAM-l induction (Fig. 3.5). This was not due to delayed signaling as the same result was obtained at later times of up to 8 h (not shown). Another striking feature of spider­venom- induced inflammation is its dependence on PMN and absence of lymphocytes and monocytes from the site of lesion. This suggested that VCAM-l expression was not upregulated up on envenomation because it is responsible for monocyte with lymphocyte recruitment. To test this we determined the surface expression of VCAM-l and found ECs exposed to spider venom failed to increase VCAM-l expression and sometimes actually decreased the small constitutive VCAM-l occasionally present on the surface (Fig. 3.5). Thus we conclude that spider venom is a poor agonist for both VCAM-l and ICAM-l expression, in addition to E-selectin. This could in part account for the lack of PMN transmigration and mononuclear cell infiltration at the site of lesion leading to a dysregulated inflammatory response. Weak NF-KB translocation induced by spider venom accounts for poor adhesion molecule expression. NF-KB is a transcription factor bound to I-KB in the cytoplasm of endothelial cells. On stimulation with TNF, NF-KB dissociates from IKB, thus exposing its nuclear localizing sequence, leading to 33 VCAM-1 - Control -TNF ICAM-1 - BRSV 104 Figure 3.5 FACS analysis of endothelial cell ICAM-l and VCAM-l surface expression. Endothelial cells were treated with buffer, TNF, or 0.3 IlI/ml Loxosceles venom for 4 h, removed from their substrate, stained with ICAM- 1 (18E3D), or VCAM-l, or control antibody. Staining of viable cells was determined as described in "Materials and Methods." These experiments are typical of two other analyses. 34 its nuclear translocation. In the nucleus, NF-KB induces transcription of E-selectin, ICAM-l, and VCAM-l genes, by binding to their promotor regions (23). Using a 32p-NF-KB consensus oligonucleotide, we determined the extent of nuclear translocation of NF-KB following treatment of endothelial cells with TNF or spider venom (Fig. 3.6). Control endothelial cells show no constitutive NF-KB in the nucleus; treatment with TNF or spider venom results in a significant increase in NF-KB binding, but the increase with TNF is at least five times greater than that obtained with spider venom. This specific binding is competed out by excess unlabeled NF-KB oligonucleotide, but not Oct-l oligonucleotide (not shown). The amount of NF-KB translocated to the nucleus parallels that of E-selectin expression seen in these cells. Thus spider venom is a poor endothelial cell agonist due to its low potency for NF-KB induction. Discussion Envenomation by the brown recluse spider can generate severe dermonecrotic lesions and may extend to hemorrhagic involvement of liver and small intestines (4,24). The size of the lesions can be large, even though these spiders only emit a tenth to a half a microliter of venom upon stimulation (4). The mechanism by which Loxosceles venom induces necrotic lesions has remained elusive. It is clear that PMN depletion completely protects against the pathologic processes induced by the venom (5), demonstrating that the proximal event is inappropriate PMN activation. Paradoxically PMN are not directly activated by the venom at relevant doses: the initiating event must lie elsewhere. Morphologic examination of tissues shows that leukocyte adherence to the capillary wall, the first observable change, occurs an hour after envenomation (J. N. Beasley, M. A. Rekow and 35 C T B C = Control T=TNF B = BRSV specific band l> Fig 3.6 Electrophoretic mobility shift assay: Nuclear extracts were prepared from endothelial cells stimulated with buffer, 1000 U Iml TNF, or 0.1 Jll/mi of BRSV for 2 h. Gel mobility shift of a consensus NF-KB oligonucleotide was performed by incubating a 32P-NF-KB probe with,.., 3 Jlg of nuclear extracts as described in "Material and Methods." Specific gel shifted bands complexed with the NF-KB probe is marked by an arrow. The data presented are one of two experiments with similar results. 36 C. R. Geren, personal communication). By 3 h, endothelial cell damage becomes apparent at the ultrastructural level (8). These events suggest that the primary action of the toxin might be on the blood vessels (25). We have confirmed this possibility: the venom of Loxosceles recIusa activates endothelial cells. One measure of this activation is that venom-activated endothelial cells bind PMN and then provoke these cells to release the lytic contents of their granules. Unusually, these bound PMN fail to become polarized, a requirement for effective directed migration, and do not emigrate through the monolayer. This endothelial cell-dependent neutrophil stasis coupled with stimulated degranulation at the intimal surface may account for the unto wards effects of envenomation. PMN normally do not interact with the vascular wall, but, in response to injury or a variety of endothelial cell- or PMN-specific agonists, they begin to roll along the vessel wall and some eventually become tightly bound. These then may exit the vascular system following their morphologic transition from round cells to migrating, polarized cells. Adhesion of PMN in response to direct agonists results from the functional upregulation of their CDll/CD18 integrins, which then bind to target molecules on endothelial cells. Adhesive interactions mediated exclusively by this mechanism are likely to occur only where shear stress from flowing blood is reduced because the strength of this interaction, at least initially, is weak (11). In contrast, activation of the endothelium results in the expression of molecules that can mediate shear-resistant interactions. In response to rapidly-acting agonists like thrombin or histamine, endothelial cells translocate P-selectin from intracellular Weibel-Palade bodies to their apical surface where it binds sialic acid-containing glycoproteins on unactivated PMN (26,27). Alternatively, endothelial cells activated by TNF, IL-l, lipopolysaccharide, and, recently, IL-3 37 (28) synthesize E-selectin and express it on their surface where it binds a sialic acid-containing glycoprotein target(s) on PMN (29). The strength of selectin­mediated tethering interactions is sufficient to result in the initial rolling interaction (30,31), but a signal to activate PMN is required for a complete inflammatory response. In the case of rapidly-activated endothelial cells this is PAF (32), whereas in cytokine-activated cells it is soluble mediators like IL-8 (15,17). Leukocyte extravasation and activation are therefore due to the expression of both a tethering selectin by activated endothelial cells, and activation of the adhesive and migratory response of the juxtaposed PMN by an endothelial cell-derived agonist. Adhesion of PMN to venom-activated endothelial cells in many ways resembled cytokine-stimulated endothelial cell-dependent adhesion. Venom exposure stimulated the synthesis and accumulation of E-selectin protein. This E-selectin was functional because we found that a panel of blocking monoclonal antibodies significantly inhibited PMN adhesion to venom­activated endothelial cells (33). Thus, Loxosceles venom stimulates aspects of endothelial cell signal transduction that result in the induction of E-selectin message synthesis and surface expression of E-selectin. These experiments also define a new endothelial cell agonist that induces responses previously only known to result from exposure to TNF, IL-1, endotoxin and, now, IL-3. The limited amount of E-selectin on the surface of venom-stimulated monolayers was presented in the context of basal levels of ICAM-1 surface expression, and an anti-ICAM-1 monoclonal antibody had no discemable effect on PMN adhesion to venom-treated monolayers. Although ICAM-1 and E-selectin can be independently regulated (34,35). Prior to this work, there was little evidence of this. One caveat here is that basal expression of ICAM-1 would cloud our detection of small increases in ICAM-1 expression; 38 however, we found another response that demonstrates that the venom does not completely mimic all of the actions of TNF. We found that under conditions in which similar levels of PMN adhesion were induced by venom and TNF, expression of IL-6 was detectable only in supernatants derived from TNF-treated monolayers. Here, where control monolayers secrete no detectable IL-6, the response to TNF and Loxosceles venom are noticeably different. Differential stimulation of endothelial cell function is also apparent when E-selectin expression is compared to elaboration of IL-S. In contrast to its weak effects on expression of this adhesion molecule, we found that elaboration of IL-S in response to venom was nearly half that induced by after 4 h of exposure and about 60% of that induced by TNF by S h (not shown). That this ratio was maintained over time shows that this truly represents differential induction rather than just an altered rate of expression. Unexpectedly, we found that venom also caused a large and early induction of GM-CSF release into the overlaying medium. Endothelial cells respond to IL-l and TNF with increased synthesis and release of GM-CSF (IS,36,37), but production is a prolonged process: TNF-evoked bioactivity is just detectable at S h with increased levels after 24 h (IS); and, our ELISA failed to detect enhanced levels of GM -CSF by 4 h of TNF stimulation. Our results show that the amount of immunoreactive GM-CSF released by 4 h venom stimulation is equivalent to that released by endothelial cells exposed to IL-l for 24 h (36): for this response, then, differential induction by these two agonists was a temporal one. Cytokine induction of E-selectin, ICAM-l, IL-8, and GM-CSF expression is complex and at least for ICAM-l (38) and GM-CSF (39), reflects both transcriptional and posttransciptional regulation. The promoter region of E­selectin, IL-S, IL-6 (40), and ICAM-l, but not GM-CSF, contains NF-lCB 39 recognition elements that are essential for enhanced transcription following cytokine stimulation. Of these, only for IL-8 is the formation of a NF-lCB complex necessary and (and in conjunction with a constitutive c/EPB transcription factor) sufficient (41) for effective transcription: E-selectin (42- 44), ICAM-1 (38,45), and IL-6 (46) require additional factors for efficient promoter activity. Additionally, regulation of E-selectin transcription may have an additional layer of complexity in that two different NF-lCB heterodimers bind the NF-lCB recognition sequence (44). E-selectin expression is also modulated under some conditions by cAMP levels (47), but we found no evidence that increased levels of cAMP accounted for the differential expression of E-selectin between TNF- and venom-treated monolayers: exposing venom-activated monolayers to 100 IlM Rp cAMP to inhibit cAMP­dependent kinase activity did not increase E-selectin expression of venom­treated monolayers (data not shown). Thus, a parsimonious explanation for our results is that venom leads to the release of NF-lCB from its cytoplasmic inhibitory complex and its translocation to the nucleus where this single event induces IL-8 transcription. In support of this, we find by electrophoretic mobility shift assays that Loxosceles venom treatment, like TNF-exposure, results in the intranuclear appearance of lCB element binding factors that appropriately shift a lCB probe (Fig. 3.6). The low production of E-selectin, IL-6, and ICAM-1 suggests that other factors essential for efficient transcription of these genes are not produced upon venom exposure. The effect of venom on GM-CSF production is less obvious given the major effect of message stability on its production (37,39) and that the identified regulatory element is a target for an unknown transcription factor. Independent control of GM-CSF and E­selectin transcription has previously been shown; an antisense oligonucleotide to the GM-CSF regulatory element blocks GM-CSF 40 expression in IL-1-stimulated endothelial cells but has no effect on E-selectin expression (36). We anticipate that the availability of an unusual agonist that differentially induces a panel of inflammatory cytokines and adhesive proteins will provide a new and valuable approach to define the intracellular signaling events in activated endothelial cells. We have identified a possible cellular and molecular basis for the PMN-dependent necrosis following envenomation by Loxosceles spiders. Prior work aimed at demonstrating a direct effect on PMN, or blood components, has not yielded a mechanism to account for the paradoxical requirement for the victim's PMN in the dermonecrosis that can follow envenomation. The potent effect of the venom on endothelial cells described here suggests that the inappropriate activation of PMN function by the venom is an indirect one. However, Loxosceles venom potently activates some, but not all, of the mechanisms that should result in the regulated extravasation of circulating PMN. The inappropriate accretion of PMN and their activation in the intravascular space suggests that this venom will be a useful tool to dissect the mechanisms that underlie normal leukocyte­endothelial cell interactions. References 1. Atkins, J. A., Wingo, C. W., Sodeman, W. A., and Flynn, J. E. (1958) Am. J. Trop. Med. Hygiene. 7, 165-184 2. Wasserman, G. S. and Anderson, P. C. (1983) J. Toxicol. Clin. Toxicol. 21, 451-472 3. Rees, R. S. and Campbell, D. S. (1989) in Management of Wilderness and Environmental Emergencies (Auerbach, P. S. and Geehr, E. C., eds) pp. 543- 548, The C.V. Mosby Company, St. Louis 4. Smith, C. W. and Micks, D. W. (1968) Am. J. Tropical. Med. Hygiene. 17, 651-657 41 5. Smith, C. W. and Micks, D. W. (1970) Lab. Invest. 22, 90-93 6. Babcock, J. L., Marmer, D. J., and Steele, R. W. (1986) Toxicon 24, 783-790 7. Majeski, J. A., Stinnett, J. D., Alexander, J. W., and Durst, G. G.,Sr. (1977) Toxicon 15, 423-427 8. Berger, R. S., Adelstein, E. H., and Anderson, P. C. (1973) J. Invest. DermatoI.61, 142-150 9. Geren, C. R., Raymond, L. N., Edwards, M., and Beasley, J. N. (1987) Fed. Proc. 46, 2291a(Abstract) 10. Zimmerman, G. A., Prescott, S. M., and McIntyre, T. M. (1992) Immunol. Today 13, 93-100 11. Lawrence, M. B., Smith, C. W., Eskin, S. G., and McIntire, L. V. (1990) Blood 75, 227-237 12. Bevilacqua, M. P. and Nelson, R. M. (1993) J. Clin. Invest. 91, 379-387 13. Erbe, D. V., Wolitzky, B. A., Presta, L. G., Norton, C. R., Ramos, R. J., Burns, D. K., Rumberger, J. M., Rao, B. N. N., Foxall, C., Brandley, B. K., and Lasky, L. A. (1992) J. Cell. Bioi. 119, 215-227 14. Lorant, D. E., Patel, K. D., McIntyre, T. M., McEver, R. P., Prescott, S. M., and Zimmerman, G. A. (1991) J. Cell. BioI. 115, 223-234 15. Rot, A. (1992) Immunol. Today 13, 291-294 16. Colditz, 1. G., Zwahlen, R. D., and Baggiolini, M. (1990) J. Leukoc. BioI. 48, 129-137 17. Huber, A. R., Kunkel, S. L., Todd, R. F.,III, and Weiss, S. J. (1991) Science 254,99-102 18. Broudy, V. C., Kaushansky, K., Segal, G. M., Harlan, J. M., and Adamson, J. W. (1986) Proc. Natl. Acad. Sci. USA. 83, 7467-7471 19. Richter, J., Andersson, T., and Olsson, 1. (1989) J. Immunol. 142, 3199-3205 20. Griffin, J. D., Spertini, 0., Ernst, T. J., Belvin, M. P., Levine, H. B., Kanakura, Y., and Tedder, T. F. (1990) J. Immunol. 145, 576-584 21. Gamble, J. R., Rand, T. H., Lopez, A. F., Clark-Lewis, I., and Vadas, M. A. (1990) Exp. Haematol. 18, 897-902 42 22. Kaufman, S. E., DiPersio, J. F., and Gasson, J. C. (1989) Exp. Haematol. 17, 800-804 23. Read, M. A., Whitley, M. Z., Williams, A. J., and Collins, T. (1994) J. Exp. Med. 179, 503-512 24. Babcock, J. L., Suber, R. L., Frith, C. H., and Geren, C. R. (1981) Toxicon 19, 463-471 25. Jolly, R. D. (1965) J. Compo Path. 75, 417-431 26. Varki, A. (1992) Curro Opin. Cell, Biol.4, 257-266 27. Moore, K. L., Stults, N. L., Diaz, S., Smith, D. F., Cummings, R. D., Varki, A., and McEver, R. P. (1992) J. Cell. Biol. 118, 445-456 28. Brizzi, M. F., Garbarino, G., Rossi, P. R., Pagliardi, G. L., Arduino, C., Avanzi, G. C., and Pegoraro, L. (1993) J. Clin. Invest. 91, 2887-2892 29. Phillips, M. L., Nudelman, E., Gaeta, F. C. A., Perez, M., Singhal, A. K., Hakomori, S. -1., and Paulson, J. C. (1990) Science 250, 1130-1135 30. Ley, K., Gaehtgens, P., Fennie, C., Singer, M. S., Lasky, L. A., and Rosen, S. D. (1991) Blood 77, 2553-2555 31. Lawrence, M. B. and Springer, T. A. (1991) Cell 65, 859-873 32. Zimmerman, G. A., Mcintyre, T. M., Mehra, M., and Prescott, S. M. (1990) J. Cell. Biol. 110, 529-540 33. Patel, K. D., Modur, V., Zimmerman, G. A., Prescott, S. M., and Mcintyre, T. M. (1994) J. Clin. Invest. 94, 631-642 34. Pober, J. S., Slowik, M. R., De Luca, L. G., and Ritchie, A. J. (1993) J. Immunol. 150, 5114-5123 35. Bradley, J. R., Johnson, D. R., and Pober, J. S. (1993) Am. J. Pathol. 142, 1598- 1609 36. Segal, G. M., Smith, T. D., Heinrich, M. C., Ey, F. S., and Bagby, G. C.,Jr. (1992) Blood 80, 609-616 37. Kaushansky, K. (1989) J. Immunol. 143, 2525-2529 38. Wertheimer, S. J., Myers, C. L., Wallace, R. W., and Parks, T. P. (1992) J. Biol. Chem. 267, 12030-12035 39. Akahane, K. and Pluznik, D. H. (1993) Exp. Haematol. 21, 878-884 43 40. Libermann, T. A. and Baltimore, D. (1990) Mol. Cell. BioI 10, 2327-2334 41. Mahe, Y., Mukaida, N., Kouji, K., Akiyama, M., Ikeda, N., Matsushima, K., and Murakami, S. (1991) J. BioI. Chern. 266, 13759-13763 42. Montgomery, K. F., Osborn, L., Hession, C., Tizard, R., Goff, D., Vassallo, C., Tarr, P. 1., Bomsztyk, K., Lobb, R., Harlan, J. M., and Pohlman, T. H. (1991) Proc. Natl. Acad. Sci. USA. 88, 6523-6527 43. van Huijsduijnen, R. H., Whelan, J., Pescini, R., Becker-Andre, M., Schenk, A. -M., and DeLamarter, J. F. (1992) J. BioI. Chern. 267, 22385-22391 44. van Huijsduijnen, R. H., Pescini, R., and DeLamarter, J. F. (1993) Nuc. Acids. Res. 21, 3711-3717 45. Voraberger, G., Schafer, R., and Stratowa, C. (1991) J. Irnrnunol. 14, 2777- 2786 46. Isshiki, H., Akira, S., Tanabe, 0., Nakajima, T., Shimamoto, T., Hirano, T., and Kishimoto, T. (1990) Mol. Cell, BioI. 10, 2757-2764 47. Parhami, F., Fang, Z. T., Fogelman, A. M., Andalibi, A., Territo, M. C., and Berliner, J. A. (1993) J. Clin. Invest. 92, 471-478 CHAPTER 4 ENDOTHELIAL CELL INFLAMMATORY RESPONSES TO TNFex: CERAMIDE-DEPENDENT AND INDEPENDENT MAP KINASE CASCADES In trod uction Tumor necrosis factor ex (TNF) signaling has been examined in the context of cell differentiation, apoptosis, and cell growth, but in vivo TNF is a mediator of systemic inflammation and immune responses against intracellular parasites (1,2). A major site of action of TNF for these effects is the vascular endothelium (3) where it induces inflammatory responses by enhancing adhesion molecule expression and cytokine secretion. The binding of TNF to the TNFR1 receptor on endothelial cells (4,5) is followed by rapid translocation of the transcription factor NF-lCB from the cytoplasm to the nucleus (6). This translocation is essential for transcription of the adhesion molecules E-selectin, ICAM-l, and VCAM-1, and the cytokines IL-6 and IL-8 (7), which all possess one or more NF-lCB promoter / enhancer elements. The regulated expression of these genes initiates the inflammatory cascade: E-selectin mediates the initial interaction of leukocytes with endothelium (8); and adherent leukocytes are activated by IL-8 to bind ICAM- 1 and other ligands on the endothelial surface. The coordinated expression of these molecules by activated endothelial cells results in tight leukocyte adhesion followed by their transmigration into tissue spaces. This juxtacrine 45 tethering and activation mechanism (9) serves to localize leukocyte activation and trafficking to areas of inflammed endothelium. The signaling pathways that couple TNF receptor activation to functional responses have remained elusive. It is now known that TNF activates Raf, which initiates the ERK MAP kinase cascade, in some cells, whereas in others it activates a parallel path(s), likely initiated by MEKK, that activates the p38 and JNK MAP kinases (reviewed in (10». The way in which these cascades might be initiated remains unknown, but recent data suggest that ceramide could be the agent that couples TNF receptor activation to at least some downstream events (11-13). TNF activates a sphingomyelinase activity in broken cell preparations (14), implying that ceramide acts high in the signaling cascade, and ceramide in some cell free systems leads to IlCB degradation and NF-lCB activation (15,16), although in other systems this does not occur (17,18). Ceramide activates an undefined ceramide-activated kinase (19); it activates a phosphatase of the ITA class (11), and it potently activates PKC ~ (20). A dominant negative form of this PKC isozyme blocks both TNF and sphingomyelinase C signaling in 3T3 cells (20). Since TNF transiently increases cellular ceramide, and increased ceramide levels activates ERK (21,22) and JNK (22) MAP kinases, and NF-lCB translocation (23), ceramide is an attractive candidate as an early, essential component of TNF signal transduction. This outline, however, has been derived only in the context of growth regulation or induction of apoptosis (24). Since it is now becoming clear that different domains of the TNF receptor signal apoptosis (25) and transcription factor activation (26,27), coupling of the receptor to inflammatory events may differ from cell-cycle-related events. Here we determined if the ceramide proposed in the context of cell growth and viability also underlies the 46 inflammatory effect of TNF on one of its major in vivo targets. We report that ceramide at high levels does activate primary cultures of human endothelial cells to synthesize some inflammatory proteins from early response genes. However the small amount of ceramide evoked by TNF was incapable of stimulating a functional response from these cells, and TNF activated MAP kinase cascades that were inaccessible to ceramide. Therefore ceramide can be a partial inflammatory stimulus for endothelial cells, but this occurs only from unphysiologic overexpression. Conversely ceramide at the low levels induced by TNF rapidly stimulates a Raf-1/ERK kinase cascade and so may couple TNF receptor activation to this pathway. Finally activation of just the ERK pathway is an incomplete stimulus, so signaling pathways induced by the TNF receptor in a ceramide-independent way are essential for the inflammatory response in endothelial cells. Results Ceramide activates endothelial cells. Ceramide is proposed as a key component of TNF signaling for certain growth and differentiation related events, but its potential as an intermediate in TNF-induced inflammatory events is undefined. Our first experiment was to determine whether ceramide was capable of eliciting inflammatory responses from endothelial cells. To do this we exposed these cells to water-soluble synthetic ceramide analogs, such as N-octanoylsphingosine (C8 ceramide). Synthetic ceramides activated endothelial cells to bind quiescent PMN, and, like the response to TNF, this could be inhibited by a blocking monoclonal antibody against E­selectin (Fig. 4.1). Since this adhesion molecule is solely expressed by activated endothelial cells (28), the synthetic ceramide acted on endothelial cells and not on PMN. Accordingly, an anti-P2 integrin antibody that blocks 47 Figure 4.1 Synthetic ceramide induces E-selectin-dependent PMN adhesion. Endothelial cells were incubated with buffer or 10 J.lM CB cerami de for 4 h, the mono layers were washed, and then incubated with buffer or 10 J.lg/ml anti-E-selectin monoclonal antibody P6E2 for 30 min. PMN adhesion was determined from microscopic images as described in "Materials and Methods." (A) Unstimulated, control incubation. (B) Unstimulated, anti-E-selectin monoclonal antibody. (C) CB-ceramide treated monolayers, control incubation. (D) CB-ceramide treated mono layers, anti-E-selectin monoclonal antibody. 48 adhesion of activated PMN failed to inhibit this endothelial cell-dependent adhesion (not shown). This stimulation also was not due to contaminating endotoxin (a potent endothelial cell agonist) as polymyxin B failed to block it (not shown). We next determined whether endogenous ceramide, derived from cellular sphingomyelin, also affected this response. Treatment of endothelial cells with sphingomyelin-specific sphingomyelinase C increased endothelial cell adhesivity for quiescent PMN, and increasing amounts of sphingomyelinase C gave enhanced levels of adhesion (Fig. 4.2). The response was sigmoidal with no apparent effect at 0.1 unit of enzyme activity. The lack of a functional response with this amount of enzyme is an important issue in light of the results presented below. Again, polymyxin B had no effect on PMN adhesion (not shown). This functional response was not immediate but developed over the same several hour period required for TNF's effects to become manifest (not shown). We assessed E-selectin surface expression by flow cytometry and found sphingomyelinase C, like TNF, activated endothelial cells from a basal state in which no E-selectin was expressed to one where significant amounts were present on the surface (Fig. 4.3 B). The amount of E-selectin induced by sphingomyelinase C treatment was always less than that expressed in response to TNF, and since this pattern obtained after 4 and 8 hr of exposure, it was not due to a markedly different rate of E-selectin expression. Nevertheless, this amount of E-selectin was sufficient to account for the increase in PMN adhesion. A blocking E-selectin monoclonal antibody significantly inhibited TNF-induced adhesion and completely locked that induced by sphingomyelinase C treatment (Fig. 4.3 A). These results show that hydrolysis of cellular sphingomyelin to ceramide is 30 i TNF 25 "........., ~0 ""-"'" C 20 .0- tCnI) 15 .c «" 10 Z :E D. 5 o 0.01 0.1 1 Sphingomyelinase (U/ml) Fig. 4.2 Cellular ceramide induces E-selectin dependent PMN adhesion. Sphingomyelinase C-induced adhesion is concentration dependent. Endothelial cells were incubated with the stated amounts of 49 sphingomyelinase (closed circles) or 1000 D/ml TNF (closed square) for 4 h, and PMN adhesion was determined. A B • Control ~ Anti-E-selectin Control TNF SMase 200 400 600 800 1000 10' Control TNF SMase 1Q2 103 intensity 104 50 Fig 4.3 PMN adhesion is E-selectin dependent. (A) Endothelial cells were incubated with buffer, 1000 U/ml TNF, or 1 U/ml sphingomyelinase for 4 h; the monolayers were washed and then incubated with buffer or 10 mg/ml anti E­selectin antibody for 30 min at 370. 111In-PMN adhesion was determined. (B) Sphingomyelinase C induces E-selectin surface expression. Endothelial cells were incubated with buffer, 1000 U/ml TNF, or 1 U/ml sphingomyelinase C for 8 h, detached from the culture dish, stained for E-selectin. an endothelial cell stimulus that induces at least one key inflammatory response of these cells. 51 Sphingomyelinase C does not reproduce all of the effects of TNF. The hypothesis that TNF signaling proceeds via an activated sphingomyelinase C predicts the ceramide produced by this enzyme should exactly mimic responses evoked by TNF. That even maximally effective amounts of sphingomyelinase were incomplete agonists for E-selectin dictated that we examine other endothelial cell responses. We found that sphingomyelinase C induced expression of the adhesion molecules VCAM-l, ICAM-l, and E­selectin, as well as the cytokines IL-6 and IL-8 (Fig. 4.4), but other than IL-6 expression, even a maximally effective amount of sphingomyelinase C was a weak agonist compared to TNF. Similarly water-soluble ceramide analogs induced expression of all these adhesion molecules and inflammatory cytokines but also were weak agonists for these responses (not shown.) Thus sphingomyelinase C stimulates endothelial cell inflammatory responses, but it appears unlikely that ceramide alone accounts for all the effect of TNF on endothelial cells. The results above suggest the 55 kDa TNF receptor must, at a minimum, activate another signaling pathway in this cell type. There is one caveat to this conclusion. Restricted pools of sphingomyelin exist, and it is postulated that the intracellular pool of sphingomyelin generates the ceramide employed in signal transduction (23,27,29). The use of extracellular sphingomyelinase C in our experiments would generate ceramide on the exterior leaflet of the plasma membrane that may not have adequate access to its signaling effector. To circumvent this limitation, we reversibly permeabilized endothelial cells with glass beads under conditions in which molecules as large as antibodies gain access to the interior of the cell (30). 52 IL-8 IL-6 VCAM-1 ICAM-1 E-selec'Un o 20 40 60 80 100 0/0 of TNF response Fig 4.4 Sphingomyelinase C differentially activates endothelial cell inflanunatory responses. Endothelial cell monolayers were treated with buffer, 1000 U / ml TNF, or 1 U / ml sphingomyelinase for 8 h before the cells were detached from culture dish and stained with monoclonal antibodies to E-select in, ICAM-l, or VCAM-l. The resulting staining pattern was analyzed by flow cytometry. In the same experiment supernatants from treated endothelial cells were collected and analyzed by ELISA for 11-6 and 11-8 secretion. The data are presented as the percentage of the TNF-induced response produced by sphingomyelinase treatment. 53 Permeabilized monolayers were treated with sphingomyelinase C, a water-soluble ceramide analog, or TNF and allowed to reseal. We found permeabilized and resealed endothelial cells still responded to TNF, as they synthesized E-selectin at levels indistinguishable from untreated monolayers, and that this procedure had only a minor stimulatory effect on monolayers exposed only to buffer (Fig. 4.5). However even when sphingomyelinase C and ceramide from the external aspect of the plasma membrane or water­soluble ceramide analogs had access to the interior of the cell, they remained weak agonists for E-selectin expression. We confirmed (Fig. 4.5) that intracellular ceramide, derived from the cell's own metabolically active pool, failed to generate significant E-selectin synthesis with PDMP, an inhibitor that raises intracellular ceramide by blocking its metabolism to glycosylceramide. We conclude the failure of ceramide to act as a full inflammatory agonist was not due to inadequate access to the cell's signaling machinery. TNF induces a weak ceramide cycle in endothelial cells. TNF stimulation can lead to enhanced levels of ceramide through a receptor stimulated sphingomyelinase(s) activity (14): resynthesis of sphingomyelin completes the ceramide cycle (11). We determined whether this also occurred in primary cultures of endothelial cells and found that TNF exposure resulted in a rapid loss of up to 40%-50% of endothelial cell sphingomyelin (Fig. 4.6). This depressed level was maintained for 1 h before returning to near basal level by 2 h. TNF correspondingly caused a small increase in ceramide levels within 10 min that continued to increase over the 2 h of the experiment (Fig. 4.7 A). However the increase was only 20% over basal levels by 30 min, with just an 80% increase following 2 h of TNF exposure. Since the loss of sphingomyelin was greater than the accumulation of ceramide, and the timing of sphingomyelin loss and ceramide accumulation differ, we infer 54 Permeabilized EC Cont SMase TNF C-2 POMP Cont SMase TNF C-2 Fig 4.5 Ceramide access to the intracellular compartment does not limit induction of E-selectin expression. Endothelial cell monolayers were reversibly permeabilized with glass beads (permeabilization efficiency >80%) to allow access of external ceramide and then immediately exposed to buffer, 1000 Vlml TNF, 1 Vlml sphingomyelinase, 10 11M C-2 ceramide, or 50 11M PDMP. After resealing and 4 h of incubation, cell homogenates were prepared and samples were immunoblotted for E-selectin. 55 Fig 4.6. Sphingomyelin levels were estimated as described in "Materials and Methods" following incubation of endothelial cells for indicated times with TNF (1000 U /ml) or sphingomyelinase (1 U /ml). Lipid levels are expressed as fold change over the level at 0 min. 56 much of the initial burst of ceramide was rapidly metabolized. We also found (Fig. 4.7 A) that TNF did not cause diacylglycerol, an activator of protein kinase C and acidic sphingomyelinase activity (23,31) to accumulate. In contrast to TNF, endothelial cells treated with sphingomyelinase C demonstrated a rapid decrease in sphingomyelin levels, equivalent to that induced by TNF treatment, and this depressed level of sphingomyelin content was maintained throughout the 2 h experiment. This resulted in a rapid and large enhancement in the cellular complement of ceramide (Fig 4.7 B): within 10 min of exposure to exogenous sphingomyelinase C there was a sustained rise in cellular ceramide up to nine times its basal level. Despite the large increases in cellular ceramide content, there was not a concomitant change in diglyceride levels, which demonstrates the substrate specificity of this enzyme. We also determined the amount of cell-associated ceramide following treatment of endothelial cell monolayers with a water-soluble ceramide analog and found a four- to five-fold increase in ceramide levels, with no increase in diglyceride levels or decrement in sphingomyelin content (not shown). We conclude that TNF induces a complete ceramide cycle in endothelial cells but that the peak ceramide accumulation is far less than that induced by exogenous sphingomyelinase C. Additionally, TNF-induced ceramide accumulation that is relevant to early signaling events was modest, being only 20-30% over basal levels. We directly examined the possibility that TNF induction of an inflammatory response proceeds through ceramide by using an amount of sphingomyelinase C that generates levels of ceramide just greater than that induced by TNF. We determined that a graded increase in cellular ceramide could be attained by varying the amount of sphingomyelinase C to which the monolayers were exposed (Fig. 4.8 A). Although this did not occur in a linear fashion, 0.1 U of enzyme produced A B 2 CI) .1.5 ! u .5 ""0 1 u. 0.5 CD != 8 u 6 .5 ~ 4 ~ 2 0 TNF 20 40 60 80 100 120 minutes Sphingomyelinase 20 40 60 80 100 120 minutes 57 ____ ceramide -0- DAG Figure 4.7 TNF and sphingomyelinase increase ceramide, but not diacylglycerol, levels in endothelial cells. Diacylglycerol and ceramide levels were measured using diacylglycerol kinase following incubation of endothelial cells for indicated times with (A) 1000 Vlml TNF or (B) 1 Vlml sphingomyelinase. Closed boxes denote ceramide levels and open squares denote diacylglycerol levels. Results are mean ± standard error. Lipid levels are expressed as fold change over the level at 0 min. 58 somewhat more ceramide accumulation than did TNF treatment. We then examined E-selectin accumulation as a function sphingomyelinase C activity and found (Fig. 4.8 B) that 0.5 and 1 U gave levels of expression that were about half that induced by TNF. However, E-selectin expression was completely undetectable when the cells were exposed to 0.1 U for sphingomyelinase C. Thus even though this amount of sphingomyelinase C increased cellular ceramide to levels somewhat greater than did TNF, this amount of ceramide was incapable of stimulating E-selectin accumulation by itself. The conclusion that must be drawn is that TNF's inflammatory effects proceeds by pathways other than, or in addition to, activation of sphingomyelinase C. TNF activates both Raf/ERK and "stress" INK, P38 kinase cascades whereas ceramide activates only the Raf IERK pathway. We sought to determine by what mechanisms TNF and ceramide activate endothelial cells and to elucidate the differences between them. The current understanding of TNF signaling is complicated in that in some systems it is Raf-dependent (26) whereas in others it is Raf-independent (32-34). Similarly cerami de leads to Raf (26) and ERK (21) activation in the "growth" pathway (Fig. 4.9) in some celis, but is considered to be part of the "stress" pathway leading to JNK activation (35) in others. First we determined if TNF activates Raf-1 in this cell type, and found that it did so (Fig. 4.10). We then sought to determine whether, and at what level, ceramide acts in the endothelial cell kinase cascade leading to inflammatory responses. We examined this by determining Raf activity in immunoprecipitates of endothelial cells treated with TNF or sphingomyelinase C for varying periods of time. We found that ceramide acted high in the kinase cascade as it activated Raf-1 (Fig. 4.10). In fact, it acted more rapidly than TNF as Raf-1 was maximally activated by just A. CD 'lf"I!) u .5 "'C '0 ~ CI) "'0 "e ~ CI) 0 Control TNF 0.1 0.5 1 SMase (U/ml) B. TNF + Fig 4.8 Graded amounts of sphingomyelinase produce graded ceramide levels that do not parallel E-selectin expression. (A) Ceramide levels. Endothelial cells were incubated with the stated amount of sphingomyelinase or 1000 Vlml TNF for 1 h followed by measurement of diacylglycerol and ceramide levels as in Fig. 5. Results are expressed as mean ± standard error of duplicate experiments. (B) E-selectin expression. Endothelial cells were incubated with the stated amounts of sphingomyelinase or \\rith 1000 Vlml TNF for 4 h before E-selectin was quantitated by immunoblot. 59 60 Sphingcomry elinase TN,F re ceptor , -::::> • Ceraride MAP kinase kinase ( Raf ? MEKK) kinases ~ , - MAP kinase kinases[ MEK MKK3 MKK4 {SEK, JNKK~ ) ~ , , MAP kinases ( ERK p38 JNK (SAPK) ) J J J I-KB PO 4 ' NF-KB translocation, transcription factor activation E-selectin VCAM-1 ICAM-1 Fig 4.9 Ceramide participation in mammalian MAP kinase cascades. The outline of MAP kinases in mammalian cells is a composite and is meant to convey a current understanding of the cascades, their relationship to one another, and alternative nomenclature. The growth factor induced pathway results in activation of erk while the p38 and JNK-l MAP kinases are activated by cellular stress. The positioning of ceramide in this outline derives from data reported here and applies to the primary cultures of human endothelial cells examined here. 61 A TNF SMase Con 0.51 21 5' 151 0.51 21 5' 151 B Fig 4.10 Raf-1 is phosphorylated and activated in a time- and concentration-dependent fashion in TNF and sphingomyelinase treated endothelial cells. (A) Kinase activity. Raf-1 was immunoprecipitated from endothelial cells treated with 1 Vlml sphingomyelinase or 1000 Vlml TNF for the stated times or was immunoprecipitated from endothelial cells treated with variable amounts of sphingomyelinase, or TNF (1000 V Iml) for 2 min (B) to determine the concentration dependence of enzyme activation. The kinase activity in the immunoprecipitates was determined at 37° for 1 h with recombinant MEK-1 as substrate. Phosphorylation of this substrate was determined following autoradiography of MEK-1 resolved during a 9% SDS-PAGE. The nonimmune control was performed with 2 min TNF cell lysate. 62 30 sec of exposure to 1 U of exogenous sphingomyelinase C. In that it must take some time for the hydrolysis of external sphingomyelin and the flip-flop of ceramide into the cell compartment, ceramide is indeed a rapid agonist for Raf-1 activation. Additionally, ceramide was a potent agonist for Raf­activation as cells treated with 0.1 U of sphingomyelinase C demonstrated a significant increase in Raf-1 activity (Fig. 4.10). Thus even low levels of ceramide, such as generated upon TNF stimulation, are rapid and effective stimuli for Raf-1 activation in endothelial cells. We next determined if rapid Raf-1 activation translated to a similarly rapid activation of downstream MAP-kinase activity. We assessed kinase activation by two approaches: we examined the phosphorylation state of the kinases by differences in their mobility during SDS-polyacrylamide electrophoresis, as phosphorylation leads to activation; and we directly assayed kinase activity by immunoprecipitation and phosphorylation of preferred substrates. We found by western blot analysis (Fig. 4.11) that both p42 (ERK1) and p44 (ERK2) MAP kinases demonstrated a change in electrophoretic mobility consistent with enhanced phosphorylation in both TNF and sphingomyelinase C-treated cells. For both agonists the mobility shift was dependent on the time of stimulation, with changes occurring as early as 5 min after agonist addition. This early change was particularly evident in cells exposed to 1 U / m1 sphingomyelinase C where near maximal mobility shifts were apparent by 5 min. An inflammatory and mitogenic phorbol ester, PMA, also altered the mobility of.both ERK1 and ERK2. In contrast, a time-dependent mobility shift of the other MAP kinases p38 and JNK-1 was detected only in TNF treated cells; these two MAP kinases did not demonstrate the band broadening associated with an enhanced phosphorylation state when material from sphingomyelinase C-treated cells SI Con T S lSI T S 301 T S PMA ERK-1nr]~ • !! • !I!!.!! ERK-2 JNK-1 ~- ....... .. p3S 63 Fig 4.11 ERK activity is stimulated by both TNF and sphingomyelinase, but JNK-1 and p38 are stimulated only by TNF. (A) MAP kinase phosphorylation state. Endothelial cells were treated with 1 Vlml sphingomyelinase (designated as S) or 1000 V Iml TNF (designated as T) for the stated times, or with 10-6 M PMA for 15 min. Celllysates were prepared, the proteins resolved by SDS-PAGE, and ERK, JNK-1and p38, (top, middle and bottom panels, respectively) detected by immunoblotting. Phosphorylation of each of the proteins results in decreased mobility on SDS-PAGE. 64 was examined. PMA behaved like sphingomyelinase C in that it specifically activated ERKl and ERK2, but failed to induce p38 or JNK phosphorylation. We confirmed this surrogate assay reflected enzyme activation by directly examining kinase activity in immunoprecipitates of celllysates, and found that the ability to phosphorylate a modified EGF receptor peptide, a substrate of ERK, increased with time of exposure to either TNF or sphingomyelinase C (Fig. 4.12 A). Again it was apparent that both sphingomyelinase C at 1 U Iml and TNF induced a rapid increase in ERK kinase activity. Moreover when we examined ERK activity as a function of sphingomyelinase C activity (Fig. 4.12 B), we found that even 0.1 U/ml of sphingomyelinase was a maximally effective stimulus for ERK activation. We then assayed phosphorylation of p38 and JNK-l substrates, ATF-2 and c­Jun, and found (Fig. 4.13) these two MAP kinases were activated in a time­dependent fashion following TNF treatment. Phosphorylation of these substrates in response to TNF was similar when compared to the ERK substrate, with JNK-l and p38 activity enhanced by 5 min of TNF exposure. In contrast, even the large amounts of ceramide generated by 1 U/ml sphingomyelinase C failed to activate either of these kinases. Thus TNF stimulates all three MAP kinases, but even supraphysiologic levels of ceramide activated only ERKl and ERK2. This shows the TNF receptor initiates signaling events that the ceramide-sensitive effector(s) does not. TNF generated ceramide is suboptimal for NF-lCB translocation. The ability to selectively activate the Raf/ERK pathway with low levels of ceramide offers the opportunity to determine if this cascade can solely be responsible for translocation of NF-KB to the nucleus, or if other TNF­initiated events account for this. The results of a electrophoretic mobility shift assay that measures the ability of nuclear proteins to bind to a labeled A B Ea. () 15000 10000 --II- TNF 5000 --e- SMase 0 0 10 20 30 minutes 8000 6000 &,4000 u 2000 O~-r~~~~~-r~~~~~ o 0.5 SMase (U/ml) 1 65 Fig 4.12 ERK kinase activity vs time. (A) ERK activity measured in celllysates employed in Fig. 4.11 after ERK immunoprecipitation and incubation with a mixture of 32p_ATP and ERK selective peptide in a 30 min assay. (B) Sphingomyelinase C dependence of ERK activity. ERK activity was determined from whole celllysates using ERK selective peptide following treatment of cells with different concentrations of sphingomyelinase for 5 min. 66 30' Con T S T S PMA JNK-1 p38 Fig 4.13 Temporal dependence of JNK-1 and p38 MAP kinase activity. The top portion of the panel shows JNK-1 activity measured in celllysates employed in Fig. 4.11 after JNK-1 immunoprecipitation and incubation with a mixture of 32P-ATP and recombinant Jun in a 30 min assay. The bottom portion of the panel shows p38 activity measured in celllysates employed in panel A after p38 immunoprecipitation and incubation with a mixture of 32P-ATP and recombinant ATF-2 in a 30 min assay. Phosphorylation of Jun and A TF-2 were determined by autoradiography following a 9% SDS-P AGE. 67 NF-lCB probe, (Fig. 4.14) show that the nuclei of quiescent endothelial cells contained a small amount of constitutive NF-lCB-like binding activity. Exposure of monolayers to TNF for 60 min alters this such that nuclei now contained large amounts of the specific band. Formation of this complex was completely inhibited by excess unlabeled NF-lCB oligomers but not by Oct-lor AP-l oligomers (not shown). Exogenous sphingomyelinase C also induced significant nuclear NF-lCB accumulation, reaching approximately half that of TNF exposed cells, but this occurred only at high sphingomyelinase C levels. Similar results were obtained by immunolocalization of the p65 NF-lCB subunit (not shown). When the amount of sphingomyelinase C was ad.justed to 0.1 U/ml, the level of NF-lCB binding activity was very much less than that of TNF-exposed cells. Thus ceramide at high levels can induce NF-lCB translocation, consistent with its proinflammatory effect, but the modest ceramide accumulation in response to TNF does not account for this. Discussion Ceramide has recently received considerable attention as a possible link between the TNF receptor and its functional responses. However, the relevant ceramide-stimulated activity or activities, the immediate downstream effector(s) of the ceramide-sensitive step(s), the potential to participate in inflammatory in addition to growth related events, or even the ability to produce physiologically functional proteins rather than message or reporter protein accumulation are all undefined. Illuminating the answers to these questions is important as TNF is a major factor in the response to bacteria both through inflammatory and immune mechanisms (1,2). The involvement of TNFRl in most of the processes associated with TNF exposure has been established, with one intracellular domain, the death TNF SMase + 1.0 0.5 0.1 68 .... NF-KB Fig 4.14 Electrophoretic mobility shift assay: Nuclear extracts were prepared from endothelial cells stimulated with buffer, 1000 V/ml TNF, or the stated concentrations of sphingomyelinase C for 2 h. Gel mobility shift of a consensus NF-lCB oligonucleotide was performed by incubating a 32P-NF-lCB probe with,.., 3 Jlg of nuclear extracts. Specific gel shifted bands complexed with the NF-lCB probe is marked by an arrow. The data presented are one of two experiments with similar results. 69 domain (25), being responsible for the apoptotic response. The regions of the receptor responsible for activation of other responses are now being elucidated, and these do not map to this particular domain. The extreme carboxy terminus of the receptor is required for NF-lCB translocation and activation of an acidic sphingomyelinase activity (27), and a recent report (26) shows a completely different, juxtamembrane region is essential for Raf-1 activation and expression from an AP-1 reporter construct. There may be a close functional relationship between one or more of these activation domains and neutral sphingomyelinase activity as TNF stimulates sphingomyelinase activity in cell-free membrane preparations (14), and TNF treatment of a variety of cell lines (17,27,31,36), and as we show here primary endothelial cell cultures cause ceramide accumulation. In some cells a more complex scheme has been proposed where in addition to the plasma membrane neutral sphingomyelinase, a lysosomal acidic sphingomyelinase is activated by diacylglycerol produced by a TNF-stimulated phospholipase C (27). This scheme would neatly explain the observation (29) that a protected, presumably intracellular, pool of ceramide is essential for modifying HL60 cell viability as this pool would be only accessible to the latter sphingomyelinase. However, fibroblasts from Nieman-Pick or I-cell disease patients that lack this lysosomal sphingomyelinase respond to TNF (37), so this dual sphingomyelinase model may not be generally applicable. In addition to the confusion regarding ceramide generation, precisely how ceramide couples to downstream functional responses is also unclear. In some experimental systems ceramide activates a 97 kDa kinase (19), or protein kinase C ~ (20), whereas in others it activates a phosphatase (38). The effects of TNF are reported to be Raf-dependent (26) in some systems and Raf­independent in others (32-34); ceramide has very recently been found (26) to 70 activate just Raf-dependent signaling. It appears then that the role of ceramide in TNF signaling depends on the experimental system and the outcome measured. We focused on a single major in vivo target of TNF and used the expression of functional protein as the readout. We asked whether ceramide was an agonist in these cells for inflammatory responses and whether it could participate in the TNF signal transduction pathway. Ceramide, either as a synthetic analog or produced by exogenous sphingomyelinase, activated the inflammatory response of primary cultures of human endothelial cells. Ceramide caused endothelial cells to bind human PMN, the first step in physiologic inflammation, through the synthesis and expression of E-selectin. Additionally, ceramide induced endothelial cell expression of inflammatory cytokines such as IL-8 and IL-6 that activate PMN and other cells that involved in an acute inflammatory response. We find a similar group of responses following exposure of endothelial cells to the sphingomyelinase D of the brown recluse spider or corynebacterium pseudotuberculosis (39). Ceramide can therefore act as an inflammatory stimulus in vitro and in vivo (unpublished, M. Feldhaus, G. Zimmerman, S. Prescott and T. McIntyre). However, activation of the endothelial cell inflammatory response occurred only after an appreciable increase in cellular ceramide content. We next investigated whether ceramide could have a role in TNF signaling in this target celL We first documented that TNF induced a sphingomyelin cycle in endothelial cells. TNF caused a 40-50% loss of sphingomyelin from endothelial cells similar to 1 U/ml of sphingomyelinase. However we found TNF enhanced ceramide levels only 20 to 30% over basal levels during the first 30 min of TNF stimulation. This was less than the enhancement caused by 0.1 U Iml sphingomyelinase C, an 71 amoWlt of enzyme that was not a stimulus for E-selectin expression or PMN adhesion. From this observation, and the fact that even high levels of cerami de accumulation failed to result in significant expression of VCAM-1 and ICAM-1, we conclude ceramide does not solely accoWlt for TNF activation of the endothelial cell inflammatory response. Furthermore, our results suggest caution in interpreting other data as the use of exogenous sphingomyelinase or ceramide analogs at inappropriate levels may lead to erroneous conclusions regarding the ability of ceramide to mimic TNF actions. We determined how signaling was effected in response to TNF and cerami de to elucidate the similarities and differences in these pathways. TNF stimulated a Raf-dependent kinase cascade and Raf-1 activation was rapid, being detectable by 30 sec of stimulation. Interestingly, MEK activation, detected as the accumulation of phosphorylated ERK1 and 2, or by activation of the ERK activity, was somewhat delayed compared to Raf-1 activation. Still, activation of these MAP kinases was apparent by 5 min of TNF stimulation. We determined whether ceramide was an agonist for this kinase cascade and fOWld that it was an excellent one. It rapidly activated Raf- 1 and did so as quickly as TNF, and this rapidity was carried through to MEK and ERK activation. Stimulation of this particular MAP kinase cascade, which is activated by growth factor receptors, is consistent with the ability of exogenous sphingomyelinase to stimulate cellular proliferation in quiescent fibroblasts (40). ERK activation in endothelial cells was exceedingly sensitive to ceramide levels such that maximal ERK activation occurred with just 0.1 U/ml of exogenous sphingomyelinase. Since this amount of enzyme generated amounts of cerami de just greater than that produced by TNF activation of an endogenous sphingomyelinase, it is possible that the early 72 stimulation of ERK activity by TNF results from the ceramide it generates. For this pathway, ceramide meets all the requirements as an intermediate in TNF signaling. It accumulates quickly enough, the amounts generated are sufficient for this role, and ceramide is a rapid activator of ERK activity. Ceramide previously has been shown to rapidly induce ERK phosphorylation in HL60 cells (21), but its effects on downstream targets appears to develop more slowly than in response to TNF (21,22). A similar inappropriate temporal relationship was obtained (26,27) in a TNF-receptor transfected B cell line where kinase activation in response to TNF was more rapid than in response to ceramide. A definitive demonstration of an essential role for ceramide in TNF activation of the Raf/ERK pathway awaits methods to either block ceramide accumulation or activation of its target effector, but at least the necessary requirements for such a role are met in endothelial cells. Our studies show that TNF also stimulated a parallel pathway(s) leading to activation of the p38 and JNK-1 MAP kinases (35,41). This pathway, which responds to osmotic shock or stress stimuli (such as UV irradiation, heat shock, or inhibition of protein synthesis), is distinct from the cascade that activates ERK (10,41). This pathway may dissolve into two paths should MKK4 specifically activate JNK-1 and MKK3 specifically activate p38 as recently postulated (41). TNF activation of JNK-1 and p38 activity was well above basal levels by 5 min but was significantly less than maximal activity at 15 min after TNF exposure. In contrast to TNF and expectations based on the literature, ceramide was incapable of activating either of the MAP kinases, p38, and JNK-1, associated with the stress-activated pathways. Chief among the reasons to expect ceramide activation of the stress pathway is its ability to activate JNK-1 activity in Hep G2 cells (35) and the recent demonstration that sphingomyelinase C stimulates Jun phosphorylation in HL60 cells (22). 73 However in human endothelial cells neither gel shifts associated with phosphorylation of p38 or JNK-l nor enhanced kinase activity against their substrates was detectable following sphingomyelinase C treatment. Thus the TNF receptor couples to p38 and JNK-l MAP kinases in endothelial cells but does so in a ceramide-independent fashion. One essential event correlated with TNF exposure is NF-lCB translocation to the nucleus following IlCB phosphorylation and degradation (42). Translocation of the p65 subunit (not shown) and formation of functional p65-p50 like complexes in the nuclei of endothelial cells occurred in response to TNF and high levels of sphingomyelinase C. However, the levels of ceramide generated by TNF was suboptimal for efficient NF-lCB translocation. Hence, activation of Raf-l alone leading to activation of ERK-l and 2 is not sufficient for this response, as found in a murine pre-B cell line (26). Additionally, since low levels of ceramide were not as effective as TNF stimulation, ceramide-independent TNF effector mechanisms promote the bulk of NF-lCB translocation. Although this might result from cooperative signals derived from p38 and JNK-l MAP kinase activation, higher levels of ceramide, which still failed to activate these MAP kinases, were able to overcome the missing event(s). These data show a variety of signals from the TNF receptor is needed for the induction of an effective inflammatory response from endothelial cells and that ceramide is only one of them. MAP kinase activation, transcription factor phosphorylation, and NF­lCB translocation lead to endothelial cell gene expression of E-selectin, ICAM- 1, VCAM-l, IL-6, and IL-8. In all these predominantly transcriptionally regulated genes, NF-lCB cooperatively associates with other transcription factors at the promotor regions to induce gene transcription (7,43). In the case of TNF induction of E-selectin, phosphorylation of ATF-2 and c-Jun 74 heterodimers by p38 and JNK-1 activate transcription in conjunction with NF-lCB (43,44), where as ceramide fails to phosphorylate ATF-2 and c-Jun, and leads to to poor gene induction. IL-6 and IL-8 are regulated transcriptionally by ERK-induced phosphorylation of C/EBP-~, in conjunction with NF-lCB (45- 47), and these are well expressed by both TNF and sphingomyelinase C. These factors were, however, insufficient for transcriptional upregulation of VCAM-1 and ICAM-l. The use of ceramide as a receptor-independent agonist thus reveals the complexity of signal integration at the promotor region in a coordinated inflammatory response. In conclusion, a role for ceramide in the expression of inflammatory gene products has been inferred from the ability of significant increases in cellular ceramide, induced by exogenous sphingomyelinase or water-soluble ceramide analogs, to mimic the effects, or some of the effects (17), of TNF on cell differentiation, induction of apoptosis, and message accumulation from reporter gene constructs. However such correlative conclusions can be misleading if significant differences in ceramide accumulation occur following the two types of stimulation. Ceramide alone-but only at high levels can induce disjointed inflammatory responses, but this is likely only relevant to bacterial sepsis by prokaryotes (e.g., C. perfingens, S. aureus, B. cereus) that express this activity as an exotoxin (48) or following envenomation by the brown recluse spider (39). Identification of a specific Raf/ERK-specific agonist, at least in this cell type, allowed us to demonstrate that modest, and early, increases in cellular ceramide following TNF treatment can activate Ras growth factor receptor-stimulated pathway. Activation of this cascade coincides with small increase in NF-lCB translocation to the nucleus, but this alone is not sufficient to induce an inflammatory response. The TNF receptor also stimulates the stress-activated 75 p38 and JNK-l kinase cascades, in a ceramide-independent fashion, and this or another ceramide-independent signal(s) to activate efficient NF-KB translocation is required for efficient induction of the inflammatory response by endothelial cells. References 1. Rothe, J., Lesslauer, W., Lotscher, H., Lang, Y., Koebel, P., Kontgen, F., Althage, A., Zinkernagel, R., Steinmetz, M., and Bluethmann, H. (1993) Nature 364, 798-802 2. Pfeffer, K., Matsuyama, T., Kundig, T. M., Wakeham, A., Kishihara, K., Shahinian, A., Wiegmann, K., Ohashi, P. S., Kronke, M., and Mak, T. W. (1993) Cell 73, 457-467 3. Pober, J. S. and Cotran, R. S. (1995) Physiol. Rev. 70, 427-451 4. Slowik, M. R., De Luca, L. G., Fiers, W., and Pober, J. S. (1993) Am. J. Pathol. 143, 1724-1730 5. Mackay, F., Loetscher, H., Stueber, D., Gehr, G., and Lesslauer, W. (1993) ]. Exp.~ed. 177,1277-1286 6. Heller, R. A. and Kronke, M. (1994) ]. BioI. Chem. 126, 5-9 7. Collins, T. (1995) Lab. Invest. 68, 499-508 8. McEver, R. P., Moore, K. L., and Cummings, R. D. (1995) ]. BioI. Chem. 270, 11025-11028 9. Zimmerman, G. A., Lorant, D. E., McIntyre, T. M., and Prescott, S. M. (1993) Am. ]. Respir. Cell. Mol. BioI. 9, 573-577 10. Cano, E. and Mahadevan, L. C. (1995) Trends. Biochem. Sci. 20, 117-123 11. Hannun, Y. A. (1994) ]. BioI. Chem. 269, 3125-3128 12. Kolesnick, R. and Golde, D. W. (1994) Cell 77, 325-328 13. Cruz, R., Sarcich, J. L., Tomasselli, A. G., and Manning, A. M. (1995) FASEB. J. 9, A127 14. Dressler, K. A., Mathias, S., and Kolesnick, R. N. (1992) Science 255, 1715- 1718 76 15. Reddy, S. A. G., Chaturvedi, M. M., Darnay, B. G., Chan, H., Higuchi, M., and Aggarwal, B. B. (1994) J. BioI. Chern. 269, 25369-25372 16. Machleidt, T., Wiegmann, K., Henkel, T., Schutze, 5., Baeuerle, P., and Kronke, M. (1994) J. BioI. Chem. 269, 13760-13765 17. Dbaibo, G., Obeid, L. M., and Hannun, Y. A. (1993) J. BioI. Chem. 268, 17762- 17766 18. Betts, J. C., Agranoff, A. B., Nabel, G. J., and Shayman, J. A. (1994) J. BioI. Chem. 269, 8455-8458 19. Liu, J., Mathias,S., Yang, Z., and Kolesnick, R. N. (1994) J. BioI. Chern. 269, 3047-3052 20. Lozano, J., Berra, E., Municio, M. M., Diaz-Meco, M. T., Dominguez, 1., Sanz, L., and Moscat, J. (1994) J. BioI. Chem. 269, 19200-19202 21. Raines, M. A., Kolesnick, R. N., and Golde, D. W. (1993) J. BioI. Chern. 268, 14572-14575 22. Saleem, A., Yuan, Z. M., Taneja, N., Rubin, E., Kufe, D. W., and Kharbanda, S. M. (1995) J. ImmunoI. 154, 4150-4156 23. DeWitt, D. L. and Smith, W. L. (1988) Proc. Natl. Acad. Sci. USA. 85, 1412- 1416 24. Pushkareva, M., Obeid, L. M., and Hannun, Y. A. (1995) ImmunoI. Today 16,294-297 25. Tartaglia, L. A., Ayres, T. M., Wong, G. H. W., and Goeddel, D. V. (1993) Cell 74, 845-853 26. Belka, C., Wiegmann, K., Adam, D., Holland, R., Neuloh, M., Herrmann, F., Kronke, M., and Brach, M. A. (1995) EMBO. J. 14, 1156-1165 27. Wiegmann, K., Schutze, 5., Machleidt, T., Witte, D., and Kronke, M. (1994) Cell 78, 1005-1015 28. Bevilacqua, M. P. and Nelson, R. M. (1993) J. Clin. Invest. 91, 379-387 29. Linardic, C. M. and Hannun, Y. A. (1994) J. BioI. Chern. 269, 23530-23537 30. Fennell, D. F., Whatley, R. E., McIntyre, T. M., Prescott, S. M., and Zimmerman, G. A. (1991) Arteriosclerosis and Thrombosis 11, 97-106 31. Schutze, 5., Potthoff, K., Machleidt, T., Berkovic, D., Wiegmann, K., and Kronke, M. (1992) Cell 71, 765-776 77 32. Westwick, J. K., Weitzel, C., Minden, A., Karin, M., and Brenner, D. A. (1994) J. Bioi. Chern. 269, 26396-26401 33. Winston, B. W., Lange-Carter, C. A., Gardner, A. M., Johnson, G. L., and Riches, D. W. H. (1995) Proc. Natl. Acad. Sci. USA. 92, 1614-1618 34. Minden, A., Lin, A., Lange-Carter, C., Derijard, B., Davis, R. J., Johnson, G. L., Karin, M., and McMahon, M. (1994) Science 266, 1719-1723 35. Kyriakis, J. M., Banerjee, P., Nikolakaki, E., Dai, T., Rubie, E. A., Ahmad, M. F., Avruch, J., and Woodgett, J. R. (1994) Nature 369, 156-160 36. Wiegmann, K., Schutze, 5., Kampen, E., Himmler, A., Machleidt, T., and Kronke, M. (1992) J. Bioi. Chern. 267, 17997-18001 37. Andrieu, N., Salvayre, R., and Levade, T. (1994) Biochern. J. 303, 341-345 38. Dobrowsky, R. T., Kamibayashi, C., Mumby, M. C., and Hannun, Y. A. (1993) J. Bioi. Chern. 268, 15523-15530 39. Patel, K. D., Modur, V., Zimmerman, G. A., Prescott, S. M., and Mcintyre, T. M. (1994) J. Clin. Invest. 94, 631-642 40. Olivera, A., Buckley, N. E., and Spiegel, S. (1992) J. Bioi. Chern. 267, 26121- 26127 41. Derijard, B., Raingeaud, J., Barrett, T., Wu, I. -H., Han, J., Ulevitch, R. J., and Davis, R. J. (1995) Science 267, 682-684 42. Brown, K., Gerstberger, 5., Carlson, L., Franzoso, G., and Siebenlist, U. (1995) Science 267, 1485-1488 43. Collins, T., Read, M. A., Neish, A. 5., Whitley, M. Z., Thanos, D., and Maniatis, T. (1995) FASEB. J. 9, 899-909 44. Kaszubska, W., van Huijsdui.jnen, R. H., Ghersa, P., DeRaemy-Schenk, A. -M., Chen, B. P. C., Hai, T., DeLamarter, J. F., and Whelan, J. (1993) Mol. Cell. Bioi. 13, 7180-7190 45. Matsusaka, T., Fujikawa, K., Nishio, Y., Mukaida, N., Matsushima, K., Kishimoto, T., and Akira, S. (1993) Proc. Natl. Acad. Sci. USA. 90, 10193-10197 46. Stein, B. and Baldwin, A. S.,Jr. (1993) Mol. Cell. Bioi. 13, 7191-7198 47. Stein, B., Cogswell, P. C., and Baldwin, A. S.,Jr. (1993) Mol. Cell. Bioi. 13, 3964-3974 78 48. Mollby, R. (1978) in Bacterial Toxins and Cell Membranes (Jeljaszewica, J. and Wadstrom, T., eds) pp. 367-424, Academic Press, London CHAPTERS ONCOSTATIN M INDUCES PROINFLAMMATORY AND MITOGENIC RESPONSES IN ENDOTHELIAL CELLS In trod uction Oncostatin M (OSM) is a 34 kDa glycosylated protein that was initially isolated based on its ability to retard growth of human melanoma and other tumor cell lines, but not normal fibroblasts (1). Sequence comparisons made with other cytokines suggest that OSM belongs to the IL-6 family of helical cytokines that share gp130 as one of the cellular receptor subunits (2). The major sources of production of OSM are macrophages and activated T-cells (1). It is also produced in copious quantities pathologically from Kaposi's sarcoma (KS), where it acts as an autocrine growth factor (3). As OSM stimulates KS cells, and KS is an endothelium-derived malignancy (4,5), it is plausible that one of the main in vivo targets of OSM is the endothelial cell. In support of this hypothesis, OSM binds endothelial cells. The number of binding sites is greater than most cell types, occupation of the receptor leads to tyrosine phosphorylation and IL-6 expression, and inhibitors of tyrosine phosphorylation block IL-6 expression (6,7). One of the major functions of endothelial cells is to localize leukocytes to a site of noxious stimulus. This is achieved by activation of endothelial cell proinfiammatory responses to induce leukocyte adhesion via up regulation of adhesion molecules E-selectin, ICAM-1, and VCAM-1 and activation via 80 cytokines IL-6 and IL-8. The expression of these genes are coordinated to bring about a regulated inflammatory response. The classical agonists for these responses are TNF, IL-1, and LPS (9). They activate endothelial cells to coexpress on their surface E-selectin, a tethering molecule, and IL-8, an activating molecule, to lead to leukocyte localization to the area of endothelial cell stimulus (10). Endothelial cells express two types of OSM receptors, a high and a low affinity receptor. The effect of OSM-induced IL-6 secretion and tyrosine phosphorylation occurs at concentrations that preferentially lead to the occupation of the high affinity receptor. However, the number of low affinity receptors greatly exceeds the number of high affinity receptors, -104,000 (Kd = 1.021 nM) vs -4000 (Kd = 0.015 nM) (11). The low affinity receptor for OSM consists of gp130 (6). The exact composition of the high affinity OSM receptor is unknown, but probably consists of gp 130, leukemia inhibitory factor (LIP) receptor ~, and an unidentified alpha subunit, characteristic of the heterotrimer receptor structure of helical cytokine family (Fig. 5.1). gp130 is a 130 kDa glycosylated transmembrane cell surface protein that binds OSM with low affinity, but as a heterotrimer with LIP receptor ex and ~ it acts as a affinity converter for high affinity LIP binding. It is also the signal transducing subunit of the IL-6 receptor and the ciliary neurotropic factor receptor (Fig 5.1) (8). To date the presence of LIF and IL-6 receptors in endothelial cells have not been documented. The effects of OSM on human endothelial cells were examined, and it was found that it induces the cell surface expression of E-selectin, ICAM-1, and VCAM-1. The expression of E-selectin and VCAM-1 occurred at high concentrations of OSM (lJ.Lg/ml), while IL-6 and ICAM-1 expression on OSM exposed endothelial cells occurred with low concentrations. The higher 81 Low aff inity receptor OSM ~ gp130 1 IL·6 OSM LIF " caFca T"c'aC~?. i lI: "'Ca,"'C ." ..... ..... ..... ..... ~ -a ~~ -a ~ ~ ~ 0 o 0 High affinity receptors Fig 5.1 gp130 participates in shared receptor complexes of the IL-6 cytokine family. gp130 is the low affinity receptor for OSM, however, with LIP receptor ~ (LIFR~) and an unidentified OSM receptor a subunit, a high affinity receptor for OSM is formed. gp130 homodimers in combination with IL-6 receptor a forms the high affinity IL-6 receptor or in combination with LIP receptor a and ~ forms the high affinity receptor for LIF (8). Thus a given response of a cell to any of the three cytokines would depend up on the local concentrations of each and the tissue expression of the different receptor subunits. 82 concentrations of OSM required for E-selectin expression correlate with low and high affinity receptor binding, whereas the lower concentrations of OSM required for 1 L-6 expression correlate to high affinity binding. In addition, OSM caused an impressive growth response in endothelial cells. Hence, OSM and IL-3 (12) describe a new functional class of proinflammatory cytokines that also act as endothelial cell mitogens. Results OSM induces PMN endothelial cell interaction via E-selectin. Inflammation upon local injection of a noxious stimulus can occur by activation of proadhesive responses in either PMN or endothelial cells alone. Activation of PMN results in the functional up regulation of CD11 / CD18 to bind to endothelial cell ICAM-1 or extracellular matrix proteins resulting in tight adhesion and chemotaxis (13-16). Treatment of PMN with concentrations of OSM similar to, or lower than that secreted by T -cells (17), resulted in no increase in adhesion to a gelatin matrix, suggesting that PMNs do not respond to OSM at these concentrations. Activation of endothelial cells can result in up regulation of their proadhesive properties (18). This results in binding of PMN at two distinct time periods: Early binding begins at about 5 min and subsides by 60 min in response endothelial cell agonists like thrombin and histamine (19,20). Delayed binding begins at about 3 h and subsides by 12 h in response to endothelial cell agonists like TNF, IL-1 and LPS (21,22). Hence to determine whether OSM activates endothelial cells to bind PMN, endothelial cell mono layers were treated with OSM at concentrations secreted by T-cells (17) for various periods of time, and PMN binding assays to the endothelium were performed. Endothelial cells thus treated bound PMN at 15 to 30 min and at 4 h (Fig. 5.2 83 A). The binding seen at 15 to 30 min was minimal in comparison to a five-fold increase in binding at 4 h. When endothelial cells were exposed to lower concentrations of OSM for 4 h, 0.1 J..Lg/ml of OSM was the lowest concentration that caused significant PMN adhesion over control (Fig. 5.2 B). Concentrations of OSM lower than this was ineffective for induction of PMN adhesion (not shown). In conclusion, OSM exposure to endothelial cells for 4 h leads to endothelial cell dependent PMN adhesion. Endotoxin contamination could not account for these responses, as boiling OSM caused a loss of all agonist activity, and coincubation of OSM with 10 J..Lg/ml polymixin B had no effect on agonist activity (not shown). Hence the increase in PMN adhesion seen may be due to endothelial expression of the adhesion molecule E-selectin, w