Multiple Signal Transduction Pathways Regulate TNF-Induced Actin Reorganization in Macrophages: Inhibition of Cdc42-Mediated Filopodium Formation by This information is current as TNF of September 28, 2021. Maikel Peppelenbosch, Elke Boone, Gareth E. Jones, S.J.H. van Deventer, Guy Haegeman, Walter Fiers, Johan Grooten and Anne J. Ridley

J Immunol 1999; 162:837-845; ; Downloaded from http://www.jimmunol.org/content/162/2/837

References This article cites 48 articles, 17 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/162/2/837.full#ref-list-1

Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists by guest on September 28, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

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

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1999 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Multiple Signal Transduction Pathways Regulate TNF-Induced Actin Reorganization in Macrophages: Inhibition of Cdc42-Mediated Filopodium Formation by TNF

Maikel Peppelenbosch,1*†‡§ Elke Boone,‡ Gareth E. Jones,¶ S.J.H. van Deventer,† Guy Haegeman,‡ Walter Fiers,‡ Johan Grooten,‡ and Anne J. Ridley*ሻ

TNF is known to regulate macrophage (M␾) migration, but the signaling pathways mediating this response have not been established. Here we report that stimulation of the 55-kDa TNF receptor (TNFR-1) induced an overall decrease in filamentous actin (F-actin), inhibited CSF-1- and Cdc42-dependent filopodium formation, and stimulated macropinocytosis. Using a panel of TNFR-1 mutants, the regions of the receptor required for each of these responses were mapped. The decrease in F-actin required both the death domain and the membrane proximal part of the receptor, whereas inhibition of filopodium formation and increased Downloaded from pinocytosis were only dependent upon a functional death domain. When the TNF-induced decrease in F-actin was inhibited using either receptor mutants or the compound D609, TNF-stimulated actin reorganization at the cell cortex became apparent. This activity was dependent upon the FAN-binding region of TNFR-1. We conclude that different domains of TNFR-1 mediate distinct changes in the M␾ cytoskeleton, and that the ability of TNF to inhibit M␾ chemotaxis may be due to decreased filopodium formation downstream of Cdc42. The Journal of Immunology, 1999, 162: 837–845. http://www.jimmunol.org/ umor necrosis factor (TNF), originally identified as an which is also observed in TNF-stimulated neutrophils, seems to be antitumoral activity, is now recognized as a major regu- mediated by an TNF-induced loss of chemotactic responses (7– T lator of inflammatory and immune responses (reviewed in 10). Indeed, we recently showed that TNF abrogated M␾ chemo- Ref. 1). The proinflammatory action of TNF can be attributed to a taxis toward CSF-1 but had no effect on CSF-1-stimulated chemo- large extent to its effects on phagocytic cells of the monocytic and kinesis (11). Therefore, TNF action in M␾ includes specific effects myeloid lineages. Macrophages (M␾)2 in particular function as on M␾ locomotion. important mediators of TNF in pathophysiology. In these cells, TNF acts by stimulating two different types of plasma mem- TNF triggers the synthesis and secretion of many proinflammatory brane-localized receptors: TNF receptor 1 (TNFR-1) and TNF re- cytokines, including TNF itself, as well as lipid mediators such as ceptor 2 (TNFR-2), both of which are expressed on M␾. Although by guest on September 28, 2021 platelet-activating-factor and prostaglandins (2, 3). M␾ activation these TNFRs are independently active in TNF signaling (12), the is itself dependent upon stimulation by the membrane-bound form large majority of TNF-induced signaling events can be mediated of TNF expressed by CD4ϩ Th1 cells (4, 5), which promotes im- solely by TNFR-1 (reviewed in Ref. 13). The addition of TNF to portant M␾ effector functions such as the production of oxygen TNFR-1 results in receptor aggregation. Subsequently, a variety of radicals and nitric oxide, the synthesis of antibiotic enzymes, and adapter molecules is recruited to the intracellular part of the re- the fusion of phagosomes with lysosomes. Thus, TNF-induced ac- ceptor, different domains apparently interacting with specific tivation is essential for the M␾-mediated destruction of pathogens adapter molecules, which in turn activate distinct signal transduc- and tumor cells (6). In addition, TNF has profound effects on tion pathways. Among the receptor associating-proteins identified phagocyte locomotion, inhibiting the migratory response of M␾ is a protein called FAN, which binds the TNFR at amino acids toward chemotactic stimuli, an effect that probably serves to retain 307–321 and activates neutral sphingomyelinase activity (14). Di- M␾ at the actual site of infection. This migration-inhibitory effect, rectly adjacent to the FAN-binding part of the receptor is the death domain (amino acids 326–413), which can bind the TNFR-1- associated death domain protein (15). The latter protein can sub- *Ludwig Institute for Cancer Research, London, United Kingdom; †Laboratory for sequently recruit other signaling proteins, including the TNFR- Experimental Internal Medicine, Academic Medical Centre, Amsterdam, The Neth- associated proteins 1 and 2 (TNFR-associated factor-1 and TNFR- ‡ erlands; Laboratory for Molecular Biology, Flemish Institute for Biotechnology, associated factor-2) (16, 17), the Fas-associated protein with death Gent, Belgium; ¶The Randall Institute, Kings College, London, United Kingdom; and ࿣Department of Biochemistry and Molecular Biology, University College of London, domain (17), receptor-interacting protein (18), and others. The pro- London, United Kingdom tein complex thus formed mediates a number of responses, includ- Received for publication May 28, 1998. Accepted for publication October 5, 1998. ing TNF-induced activation of NF-␬B, c-Jun N-terminal kinase The costs of publication of this article were defrayed in part by the payment of page (stress-activated protein kinase), caspases, and apoptosis (19). Fi- charges. This article must therefore be hereby marked advertisement in accordance nally, the membrane proximal region of the receptor (amino acids with 18 U.S.C. Section 1734 solely to indicate this fact. 205–258) binds and, upon receptor activation, stimulates a phos- 1 Address correspondence and reprint requests to Dr. Maikel Peppelenbosch, Labo- ratory for Experimental Internal Medicine, G2-130, Academic Medical Centre, phatidylinositol-4-phosphate 5-kinase (20). In addition, this part of Meibergdreef 9, NL-1105 AZ Amsterdam, The Netherlands. E-mail address: the receptor together with the death domain induces nitric oxide [email protected] synthase (21). 2 Abbreviations used in this paper: M␾, macrophage(s); TNFR, TNF receptor; TPA, The role of the different TNFR-1 domains and associated signal 12-O-tetradecanoylphorbol-13-acetate; PKC, protein kinase C; MAPK, mitogen-ac- ␾ tivated protein kinase; PAK, p21-activated kinase; TRITC, tetramethylrhodamine B transduction pathways in mediating the effects of TNF on M isothiocyanate. migration has not been investigated. However, a group of proteins

Copyright © 1999 by The American Association of Immunologists 0022-1767/99/$02.00 838 TNF SIGNAL TRANSDUCTION TO THE CYTOSKELETON IN M␾ known as the Rho family of small GTPases has been implicated in England Biolabs, Beverly, MA), or 100 ng/ml human recombinant TNF regulating cell migration. Members of this family, including (Genzyme, Cambridge, MA) for 5 min, and subsequently stimulated with Cdc42, Rac, and Rho, play a major role in remodeling the actin 100 ng/ml human recombinant CSF-1 (Chiron Corporation, Emeryville, CA) for 10 min (J774.2 cells), 15 min (Bac1.2F5 cells and 4-4 cells), or 20 cytoskeleton in response to external stimuli (reviewed in Refs. 22 min (P388D1 cells), or stimulated for 30 min with 1000 infectious units/ml and 23). Cdc42 specifically induces the formation of filopodia in murine TNF (4-4 cells; murine TNF was produced in Escherichia coli, Swiss 3T3 cells (24, 25), whereas Rac and Rho regulate lamelli- purified to Ն99% homogeneity in our laboratory, and contained 4 ng of podium and stress fiber formation, respectively (26–27). Recently, endotoxin/mg protein). For testing the action of mutated human TNFRs, 4-4 cells (107 cells in 1 ml of RPMI 1640 containing 5% FCS) were trans- we demonstrated that Rho family members have similar activities fected with a total of 5 ␮g of cDNA by electroporation (300 V/950 ␮F/10 in M␾: Cdc42 mediates filopodium formation, whereas Rac reg- ms), and seeded on coverslips. After 24 h, cells were stimulated for 30 min ulates lamellipodium production (28). As changes in cell locomo- with the htr-1 mAb (100 ng/ml; a gift of Dr. M. Brockhaus, Hoffman-La tion are dependent upon controlled remodeling of the actin cy- Roche, Basel, Switzerland), which clusters and thus activates the trans- toskeleton, it is generally assumed that Rho family members have fected human TNFRs but not the endogenous murine receptors (31, 32). To investigate pinocytosis, fluorescently labeled BSA (prepared using Flu- an important function in directing cell motility (29). Therefore, the orlink; Amersham, Arlington Heights, IL) was added to the culture me- TNF-induced inhibition of M␾ chemotaxis might well be brought dium in some experiments together with htr-1 Ab. about by interference with either the activation of Rho family pro- After stimulation, cells were fixed in 3.5% formaldehyde/PBS (v/v). For teins or downstream signal transduction to the actin cytoskeleton. visualization of actin filaments, cells were permeabilized in 0.1% Triton X-100/PBS (v/v), blocked in 0.5% BSA (Sigma, St. Louis, MO)/PBS (w/v) Little is known of the signaling pathways regulating TNF-in- for 45 min, and stained with 80 ng/ml tetramethylrhodamine B isothiocya- duced actin reorganization in any cell type; in particular, the effects nate (TRITC)-conjugated phalloidin (Sigma) in PBS for 1 h. For immu- of TNF on the actin cytoskeleton of M␾ have remained unex- nostaining, cells were blocked for1hinPBScontaining 10% FCS (v/v), Downloaded from plored. To obtain more insight into the molecular mechanisms by 5% milk powder (w/v), 0.5% BSA (w/v), and 0.1% Triton X-100 (v/v). which TNF controls M␾ migration, we have investigated the ef- Subsequently, cells were incubated in PBS containing 1% BSA (w/v) and ␾ 0.1% Triton X-100 (v/v) supplemented with a 1/2000 dilution of the htr-9 fects of TNF on the M actin cytoskeleton and on the actin reor- Ab (for the detection of transfected cells; the htr-9 Ab was also a kind gift ganization induced by CSF-1 and Rho family proteins. We have of Dr. M. Brockhaus). Primary Ab binding was detected by incubation for subsequently used TNFR-1 mutants to determine which portions 1 h with a 1/500 dilution of FITC-conjugated rabbit anti-mouse IgG (Jack- son ImmunoResearch Laboratories, West Grove, PA) in PBS containing of the receptor are responsible for signaling to the actin cytoskel- http://www.jimmunol.org/ ␾ 1% BSA (w/v) and 0.1% Triton X-100 (v/v). Cell images were generated eton in M . using confocal laser scanning microscopy as described previously (33, 34). Materials and Methods Microinjection Cell culture Protein expression, purification, and a determination of the active protein concentration were performed as described previously (26, 27). Microin- P388D1 cells, 4-4 cells, and J774 cells were cultured in RPMI 1640 me- jection were performed in RPMI 1640 without serum at 37°C. At least 30 dium (Life Technologies, Gaithersburg, MD) with 10% FCS according to cells were microinjected for every experimental condition. Where relevant, routine procedures. For microinjection and cell-stimulation experiments, 100 ng/ml TNF or 5 mM of C2-ceramide was added to this medium. Cells cells were seeded on 13-mm coverslips, switched to RPMI 1640 without were microinjected with 50 mM of Tris buffer (pH 7.3 at 37°C) containing FCS after 24 h, and microinjected 24 h later. In cell-stimulation experi- ␮ by guest on September 28, 2021 100 mM NaCl, 5 mM MgCl2, 150–700 g of recombinant protein/ml, and ments, cells were washed with RPMI 1640 without FCS at 24 h after 800 ␮g/ml rat IgG. After microinjection and, if appropriate, stimulation seeding and maintained without FCS for 48 h. Experiments were per- with 100 ng/ml CSF-1, cells were fixed in 3.5% formaldehyde/PBS (v/v) formed in the same medium. Bac1.2F5 cells were cultured and passaged as for Ն30 min. Subsequently, cells were blocked for1hinPBScontaining described previously (11). For experiments (unless otherwise indicated), 10% FCS (v/v), 5% milk powder (w/v), 0.5% BSA (w/v), and 0.1% Triton cells were starved overnight in RPMI 1640 without serum. Experiments X-100 (v/v) and stained with 80 ng/ml TRITC phalloidin and a 1/500 were performed in the same medium. dilution of FITC-conjugated rabbit anti-rat IgG (Jackson Laboratories) in Construction of hTNF-R55 mutants PBS containing 1% BSA (w/v) and 0.1% Triton X-100 (v/v). Microin- jected cells were identified using the anti-mouse IgG signal, and actin The cDNA of human TNF-R55 (generously provided by Drs. W. Lesslauer filaments were analyzed as described above. and H. Loetscher, Hoffmann-La Roche, Basel, Switzerland) was cloned Uptake of [3H]sucrose into the eukaryotic expression vector pCDM8 (Invitrogen, San Diego, CA). All other mutants were also cloned into this vector. The deletion mutant Where appropriate, cells were preincubated for 16 h with 100 ng/ml 12- ⌬202–304 was created by cutting the cDNA of hTNF-R55 with EarI and O-tetradecanoylphorbol-13-acetate (TPA). Pinocytosis was assayed in Eco57I, blunting both sticky ends with Klenow polymerase and T4-exo- RPMI 1640 containing 0.5 ␮Ci of [3H]sucrose (Amersham). After the in- nuclease, respectively, and religating both fragments. This process gener- dicated incubation times, cells were washed five times with ice-cold RPMI ated a receptor mutant with the following cDNA sequence at the joint 1640 and lysed in 1% SDS (w/v). Subsequently, cellular sucrose uptake fragments: 5Ј-CTC CTC TTC ATT GCC ATC CCC AAC-3Ј; the nucle- was determined by scintillation counting. Each datapoint represents the otides filled in with Klenow polymeraseides are shown in bold. The re- average of at least three independent determinations. sulting protein lacks amino acids 203–303 of human TNF-R55, but still contains the FAN-binding domain (14). Mutagenesis of leucine 351 to Results alanine (L351A) was performed according to the instructions provided TNF acts via TNFR-1 to induce a decrease in polymerized actin with the site-directed mutagenesis kit (Clontech, Palo Alto, CA). The fol- ␾ lowing mutator oligo was used: 5Ј-GGT CGC TCG CCC CTA GGC GCC in mouse M GCA C-3Ј. The mutant receptor containing only the death domain (without The effects of TNF on actin organization in M␾ were investigated the FAN-binding site) was created as follows: a cDNA fragment containing in four different mouse M␾ cell lines to ensure that responses were the death domain of human TNF-R55 preceded by a 15-aa linker (30) was ligated to the extracellular and transmembrane part of human TNF-R55 and widely observed and not specific to only one cell line. To distin- again generated by cutting human TNF-R55 with EarI and treating the guish between the effects mediated by TNFR-1 and TNFR-2, re- fragment with Klenow polymerase. All mutants were verified by sequenc- sponses to both human and mouse TNF were investigated, as hu- ing. man TNF can only stimulate TNFR-1 in mouse cells (13). Cell stimulation, immunocytochemistry, and microscopy Unstimulated, serum-starved, Bac1.2F5 M␾ were rounded, and displayed relatively high levels of polymerized actin within the Experiments were performed at 37°C, each experimental condition was tested on three different coverslips, and experiments were repeated several cell interior (Fig. 1A). Serum-starved J774.2 cells were more times (2–5). Where appropriate, cells were preincubated with 0.3 ␮Mof spread and elongated compared with Bac1.2F5 cells (Fig. 1D). SB203580 for 120 min, 4 ␮M of cell permeable ceramide for 30 min (New Unstimulated P388D1 cells exhibited an elongated phenotype with The Journal of Immunology 839 Downloaded from

FIGURE 1. TNF induces actin reorganization in M␾. Bac1.2F5 (A–C), J774.2 (D–F), and P388D1 (G–I) mouse M␾ cells were stimulated for 30 min with control vehicle (A, D, and G) or 100 ng/ml human TNF (B, E, and H) or preincubated for 1 h with 20 ␮g/ml D609 and then stimulated with TNF http://www.jimmunol.org/ (C, F, and I). Subsequently, cells were fixed and F-actin was visualized using TRITC phalloidin. The bar represents 20 ␮m. occasional microspikes; the majority of filamentous actin (F-actin) ing through the transfected receptors without activating endoge- was in the cell cortex (Fig. 1G). Upon stimulation with either hu- nous murine TNFRs (31, 32). After fixation, the actin cytoskeleton man or mouse TNF, all three M␾ cell types displayed a dramatic of transfected cells was compared with untransfected neighboring decrease in TRITC-labeled phalloidin staining. In Bac1.2F5 cells, cells by double-staining the cells with phalloidin and an Ab spe- this effect was already observed after 5 min, was maximal after 30 cific for human TNFR-1. Activation of full-length human TNFR-1 min, persisted for Ն2 h, and was somewhat more pronounced at in transfected cells led to a decrease in phalloidin staining com- by guest on September 28, 2021 the cell periphery compared with the perinuclear region (Fig. 1B). pared with untransfected cells (Fig. 2, A and B), confirming that In J774.2 M␾ cells, a decrease in F-actin was first apparent at 15 activation of TNFR-1 is sufficient to induce the decrease in F-actin. min after TNF addition; a clear effect was observed at 30 min (Fig. Subsequently, we tested the effect of a TNFR-1 containing the 1E). The decrease in F-actin in P388D1 cells principally involved L351A mutation (homologous to the lpr mutation in the Fas Ag; cortical actin structures (Fig. 1H) and followed a similar time 36), which renders the death domain inactive. Stimulation of this course to that observed in J774.2 cells. In 4-4 cells, TNF caused a receptor actually induced an increase rather than a decrease in response that was similar to that observed in Bac1.2F5 cells (data F-actin (Fig. 1, I and J); however, in contrast to the effect of D609, not shown). We conclude that TNF causes a reduction in F-actin in this increase was not limited to the cell cortex (possibly because M␾ and that TNFR-1 stimulation is sufficient for this response. the higher expression of transfected TNFR compared with the en- dogenous receptor). Surprisingly, activation of a TNFR lacking the Membrane proximal region and death domain of TNFR-1 are membrane proximal part of the receptor but containing a func- both essential for TNF-induced loss of F-actin tional death domain and a FAN-binding domain (deletion of amino To gain insight into the signaling mechanisms mediating TNF- acids 202–304) also resulted in increased levels of F-actin. A re- dependent actin reorganization, we first investigated the effects of ceptor mutant lacking both the membrane proximal part and Fan- D609. This compound inhibits a variety of death domain-depen- binding site but containing a functional death domain (see Mate- dent responses to TNF (35). Treating P388D1 cells or J774.2 cells rials and Methods) did not induce any detectable actin for 1 h with 20 ␮g/ml D609 did not detectably alter the actin reorganization (Fig. 2, E and F). Therefore, loss of polymerized organization of these cells (data not shown). However, after sub- actin in response to TNFR-1 activation requires both the mem- sequent stimulation with TNF, the TNF-induced decrease in F- brane proximal part of the receptor and a functional death domain, actin was not observed; an accumulation of F-actin at the cell whereas the FAN-binding site appears to mediate a TNF-induced cortex was observed instead (Fig. 1, C, F, and I). These data sug- increase in F-actin. gest an involvement of the death domain-dependent phosphatidyl choline-specific phospholipase C (35) in the loss of F-actin in re- TNFR-1 stimulates macropinocytosis through the death domain sponse to TNF and also indicate that TNF is able to induce actin Macropinocytosis can be stimulated in M␾ by a number of stimuli, reorganization at the cortex when the overall decrease in F-actin is including phorbol esters, and involves actin reorganization (37). prevented. TNF induced the accumulation of vesicles in M␾ (Fig. 1B), an To investigate further the signaling mechanisms underlying effect that was sensitive to D609 (Fig. 1C). To test whether these TNF-induced actin reorganization, we transiently transfected mu- vesicles were a result of a TNF-induced increase in macropinocy- tants of human TNFR-1 into mouse 4-4 M␾. Subsequent clustering tosis, we transfected 4-4 M␾ with human TNFR-1 and compared of these transfected receptors, using the Ab htr-1, initiated signal- the uptake of fluorescently labeled BSA by transfected cells with 840 TNF SIGNAL TRANSDUCTION TO THE CYTOSKELETON IN M␾ Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 2. Effects of different TNFR-1 domain mutants on actin reorganization and pinocytosis. Full-length human TNFR-1 (A and E), human TNFR-1 with a functional death domain but without the membrane proximal region (B and F), human TNFR-1 with a functional death domain but without the membrane proximal region and the FAN-binding site (C and G), or full-length human TNFR-1 containing a truncated death domain (D and H) were transiently transfected into mouse 4-4 M␾. Subsequent clustering of these transfected receptors was induced with the Ab htr-1. After fixation, transfected cells were identified by staining with an Ab specific for human TNFR-1. To investigate actin organization, cells were costained with TRITC-labeled phalloidin (E–H), whereas pinocytosis was assessed by uptake of fluorescently labeled BSA (A–D). Transfected cells are indicated with arrows. The bar represents 40 ␮m. that of untransfected cells. After stimulation with the Ab htr-1, B and C). Subsequently, we performed experiments to establish the human TNFR-1-expressing cells showed a considerably higher up- domains of TNFR-1 necessary for this effect of TNFR-1. A recep- take of fluorescent BSA compared with neighboring cells (Fig. 2, tor mutant in which both the membrane proximal part and the The Journal of Immunology 841 Downloaded from http://www.jimmunol.org/ FIGURE 3. TNF stimulates macropinocytosis in J774.2 cells. A, Time FIGURE 4. Prolonged incubation of M␾ with TPA inhibits pinocytosis. course of TNF-stimulated [3H]sucrose uptake. Cell were stimulated with Experiments were performed with growing J774.2 M␾. A, Cells were stim- vehicle (f and dotted line) or with 100 ng/ml mouse TNF (F and solid ulated with 100 ng/ml TPA (Œ and solid line) or treated with TPA for 16 h line) for the times indicated, and sucrose uptake was determined. B, Dose- and then restimulated with TPA (f and dotted line). Untreated cells are dependent stimulation with mouse TNF. Cells were stimulated with the also shown (E and dashed line). [3H]sucrose uptake was determined. B, TNF concentrations indicated; [3H]sucrose uptake was determined after [3H]sucrose uptake induced by 100 ng/ml mouse TNF in control cells (F 60 min. and solid line) and in cells treated for 16 h with 100 ng/ml TPA (f and dotted line).

FAN-binding site were deleted was still able to mediate an in- by guest on September 28, 2021 creased uptake of fluorescent BSA (Fig. 2, G and H), whereas suggest that PKC or another phorbol ester-binding protein medi- TNFR-1 containing the L351N mutation (and thus having an in- ates TNF-stimulated macropinocytosis. active death domain) was unable to induce an increased BSA up- take (Fig. 2, K and L). We conclude that increased macropinocy- CSF-1-induced filopodium formation is inhibited by TNF tosis in response to TNF is solely mediated by the death domain of TNF can inhibit phagocyte chemotaxis (39), and we have specif- the receptor. ically shown that TNF inhibits the chemotaxis of Bac1.2F5 cells toward CSF-1 but has no effect on CSF-1-induced chemokinesis Phorbol esters regulate TNF-induced macropinocytosis (11). To investigate the mechanisms underlying the ability of TNF To quantify the extent of TNF-induced macropinocytosis, the pi- to inhibit chemotaxis, we decided to characterize the effects of nocytotic activity in J774.2 M␾ was assayed directly by measuring TNF on CSF-1-induced actin remodeling. CSF-1 induces rapid [3H]sucrose uptake. Incubating cells with 100 ng/ml mouse TNF actin reorganization in Bac1.2F5 cells: it stimulates the formation caused an increase in the rate of fluid uptake, and this effect per- of filopodia, lamellipodia, and membrane ruffles (28). A similar sisted for Ն1 h (Fig. 3A). Subsequent dose-response experiments response to CSF-1 is observed in J774.2, P388D1, and 4.4 M␾ established that an effect of TNF on fluid uptake was already ob- (Fig. 5, A, E, and I). served at TNF concentrations in excess of 10 ng/ml, whereas the In TNF-treated M␾, CSF-1 was still able to induce substantial response was maximal at 50 ng/ml (Fig. 3B). The inhibition of reorganization of the actin cytoskeleton; however, filopodium for- TNF-induced macropinocytosis by D609 suggested that stimula- mation was almost completely inhibited. In contrast, CSF-1-in- tion of phosphatidylcholine-specific phospholipase C could be in- duced lamellipodium production and ruffling were enhanced by volved in this response (35). As phospholipase C activation leads TNF pretreatment in Bac1.2F5 (Fig. 5C), J774.2 (Fig. 5G), and to the production of diacylglycerol and to the activation of several P388D1 cells (Fig. 5K). Similar effects were observed when TNF protein kinase C (PKC) enzymes (38) and as PKCs are known to and CSF-1 were added at the same time (data not shown). We regulate macropinocytosis (37), the effect of the PKC activator conclude that TNF specifically inhibits CSF-1-induced filopodium TPA on macropinocytosis was investigated. The addition of 100 formation, whereas lamellipodium formation and ruffling are not ng/ml TPA potently stimulated fluid uptake in J774.2 M␾ (Fig. sensitive to TNF. 4A). To investigate whether PKC is involved in the TNF response, cells were treated with TPA for 16 h to down-regulate phorbol TNF inhibits filopodium formation downstream of Cdc42 ester-responsive PKC isoforms. A subsequent application of TPA As Cdc42 mediates CSF-1-induced filopodium extension in M␾ was no longer able to stimulate macropinocytosis (Fig. 4A). Inter- (28), the question arises as to whether TNF inhibits Cdc42 acti- estingly, TNF-induced macropinocytosis was also inhibited after vation or whether it interferes with the signaling pathway leading TPA pretreatment (Fig. 4B). Taken together, these results strongly from activated Cdc42 to the cytoskeleton. To distinguish between 842 TNF SIGNAL TRANSDUCTION TO THE CYTOSKELETON IN M␾

FIGURE 5. Effects of TNF on CSF-1-induced actin reorganization in M␾. Bac1.2F5 (A–D), J774.2 (E–H), and P388D1 (I–L) mouse M␾ cells were Downloaded from stimulated with vehicle (A, E, and I), 100 ng/ml human CSF-1 (B, F, and J), or 100 ng/ml human TNF and CSF-1 (C, G, and K), or preincubated with 0.3 SB203580 and then stimulated with TNF and CSF-1 (D, H, and L) as described in Materials and Methods. Subsequently, cells were fixed and F-actin was visualized using TRITC phalloidin. The bar represents 20 ␮m.

these two possibilities, J774.2 M␾ were microinjected with Cdc42 and Cdc42-dependent filopodium production (Fig. 6C). We con- V12 (a constitutively active mutant of Cdc42) in the presence or clude that the inhibitory effect of TNF on filopodium production is http://www.jimmunol.org/ absence of TNF. As expected, Cdc42 V12 induced a large increase mediated at least in part by p38 MAPK. in filopodium formation (Fig. 6A), but this response was com- pletely inhibited by TNF (Fig. 6B). In contrast, TNF did not in- fluence the actin reorganization induced by a microinjection of Rac Discussion V12 and Rho V14. Therefore, TNF specifically interferes with the ␾ signal transduction pathway leading from activated Cdc42 to filop- TNF stimulates a number of responses in M , but its effects on the ␾ odium production. M actin cytoskeleton have not been studied in detail. We report here that TNF induces a decrease in polymerized actin and an by guest on September 28, 2021 TNF-mediated inhibition of filopodium formation requires p38 increase in macropinocytosis and inhibits Cdc42-mediated filopo- mitogen-activated protein kinase (MAPK) dium extension. In addition, when the TNF-induced decrease in TNF is known to activate members of the MAPK family of pro- F-actin is inhibited using D609 or receptor mutants, a distinct type teins (40). Therefore, we investigated the roles of p42/44 MAPK of actin reorganization involving an increase in F-actin is observed and p38 MAPK in TNF-induced actin reorganization using the in response to TNF. For all of these effects of TNF, activation of MAPK/extracellular signal-regulated kinase inhibitor PD098059 TNFR-1 was sufficient, as specific activation of transfected (41) and SB203580, a p38 MAPK inhibitor (42). PD098059 or TNFR-1 by the htr-1 Ab induced all of the responses observed SB203580 did not alter the TNF-induced decrease in F-actin or with TNF. Similarly, these responses to TNF could be elicited by macropinocytosis (Table I). However, SB203580 (Fig. 5) but not adding human TNF to mouse M␾, which will result in the activa- PD098059 (data not shown) was able to completely abrogate the tion of only mouse TNFR-1 and not mouse TNFR-2 (13). Conse- inhibitory effect of TNF on CSF-1-induced filopodium production quently, TNFR-2 does not appear to be required for TNF-induced in P388D1 cells and partial restored the response in Bac1.2F5 and cytoskeletal reorganization in mouse M␾. J774.2 M␾. Furthermore, treatment of cells with 4 ␮MofC2- Transfection of specific TNFR-1 mutants allowed us to pinpoint ceramide, which potently activates p38 MAPK, reduced CSF-1- the elements in the receptor mediating the different effects of TNF

FIGURE 6. Effects of TNF on Cdc42-induced actin reorganization in J774.2 M␾. Cells were injected with Cdc42 V12 proteins in RPMI 1640 (A), RPMI 1640 containing 100 ng/ml TNF (B), or RPMI 1640 containing 4 ␮M of C2-ceramide (C). Subsequently, cells were fixed and F-actin was visualized using TRITC phalloidin and confocal laser scanning microscopy. The bar represents 20 ␮m. The Journal of Immunology 843

Table I. Effects of PD098059 and SB203580 on TNF-induced vious results imply that it mediates an increase, rather than de- [3H]sucrose uptakea crease, in F-actin (44). The death domain, in turn, stimulates a caspase-mediated proteolytic cleavage of p21-activated kinase 2 Stimulus cpm (Ϯ SE) (PAK2), resulting in a constitutively active PAK2 fragment; PAK1 Control 368 Ϯ 41 is not affected (45). PAKs interact with Rac and Cdc42 and re- TNF 694 Ϯ 103 portedly play a role in regulating actin organization in several cell TNF ϩ 0.3 ␮M SB203580 598 Ϯ 48 types (46). Therefore, it will be interesting to investigate the action TNF ϩ 1 ␮M PD098059 755 Ϯ 152 of phosphatidylinositol-4-phosphate 5-kinase and PAK2 on the a Cells were either left unstimulated or stimulated with TNF in the presence or M␾ actin cytoskeleton, to determine their role in TNF-induced absence of PD098059 and SB203580 as described in Materials and Methods. Pino- cytoskeletal remodeling, and to assess possible interplay between cytosis was determined by [3H]sucrose uptake. the effects of these proteins. When the TNF-induced decrease in F-actin was inhibited by the on the actin cytoskeleton (Fig. 7). We observed that for the TNF- compound D609 or by activating TNFR mutants lacking either the induced decrease in F-actin, both the membrane proximal region of membrane proximal region or the death domain, a TNF-induced TNFR-1 and the death domain were essential. The requirement for increase in F-actin was observed. This finding suggests that the these two domains is unusual and has not been observed for most signal inducing the decrease in F-actin is sufficiently strong enough other TNF-induced responses, for which either the death domain or to override other signals from TNFR-1 to the actin cytoskeleton, the FAN-binding site is sufficient; interestingly, TNF-induced in- but that when this signal is reduced, less dominant signals can duction of nitric oxide synthase also requires these two domains of induce a distinct type of actin reorganization. One possibility is Downloaded from TNFR-1 (21). In addition, L929 cells transfected with a human that TNF activates a protein that enhances actin filament turnover, TNFR-1 lacking the membrane proximal region display markedly such as ADF/cofilin (47), which then rapidly depolymerizes any delayed cell death in response to the htr-1 Ab as compared with the actin-containing structures produced in response to other TNF- full-length receptor, although NF-␬B activation was not affected activated signals. According to this model, these other structures (43). Apparently, the membrane proximal part of the receptor ex- would be more stable when the rate of actin depolymerization is erts important cooperative functions with the death domain of the decreased. The TNF-induced increase in F-actin was mechanisti- http://www.jimmunol.org/ receptor, and may be more important in TNF signaling than pre- cally distinct from the decrease in F-actin, as it was dependent viously thought. upon the FAN-binding site of the receptor. The signaling pathways Both the membrane proximal region and the death domain ac- downstream of FAN are still unclear but include activation of neu- tivate signaling proteins that could participate in the TNF-depen- tral sphingomyelinase activity, generating ceramide (14). How- dent decrease in F-actin. Recently, it has been reported that the ever, an application of cell-permeable ceramides did not induce an membrane proximal part of the receptor binds phosphatidylinosi- increase in F-actin (Fig. 6), indicating that ceramide production is tol-4-phosphate 5-kinase, and that this enzyme is activated upon not sufficient for this response. The signaling mechanisms by receptor stimulation (20). As phosphatidylinositides regulate the which the FAN-binding site mediates actin reorganization in M␾ by guest on September 28, 2021 activity of many actin-organizing proteins, this enzyme may par- may well involve Rho family proteins; therefore, we plan to in- ticipate in the actin reorganization induced by TNF, although pre- vestigate this possibility. We observed that TNF was a potent stimulator of macropino- cytosis. For this effect, a functional death domain was sufficient, and D609 also inhibited the macropinocytosis induced by TNF. The molecular target for D609 is apparently a TNF-activated phos- phatidylcholine-specific phospholipase C, which leads to PKC ac- tivation through diacylglycerol production (35). Phorbol esters, which mimic the action of diacylglycerol and activate PKC, po- tently stimulate macropinocytosis in M␾ (37); down-regulation of PKC by prolonged treatment with phorbol esters impaired TNF- induced pinocytosis. These results suggest that a phorbol ester- sensitive isoform of PKC is required for TNF-induced macropi- nocytosis. However, D609 also inhibited phorbol ester-dependent pinocytosis in M␾ (M.P.P., unpublished observations), suggesting that this compound acts downstream rather than upstream of di- acylglycerol release to inhibit pinocytosis and consequently has multiple cellular targets. It is also possible that another phorbol ester-binding protein apart from PKC is involved in the macropi- nocytic response (for example, a member of the chimerin protein family that acts as a GTPase activating protein for Rac and Cdc42) (48). Indeed, expression of activated Rac has been shown to stim- ulate macropinocytosis in fibroblasts (27). Macropinocytosis can act as a nonspecific mechanism for Ag capture, leading to Ag presentation and T cell activation (37). TNF is a potent inducer of Ag presentation in M␾ (6), and TNF-stimulated macropinocytosis may therefore be important for this response; current experiments address this possibility. FIGURE 7. Schematic depiction of the different domains in TNFR-1, Alternatively, TNF-induced macropinocytosis may serve to the roles of each domain in mediating actin reorganization in M␾, and the translocate activated receptors into intracellular compartments, interaction of these domains with CSF-1-induced signaling. as it has been shown that a number of TNF effects require TNFR 844 TNF SIGNAL TRANSDUCTION TO THE CYTOSKELETON IN M␾ internalization and the subsequent acidification of the endoso- References mal compartment (35), although this scheme makes the effects 1. Beyaert, R., and W. Fiers. 1994. Molecular mechanisms of tumor necrosis factor- of membrane-bound TNF, which is not internalized, difficult induced cytotoxicity: what we do understand and what we do not. FEBS Lett. to explain. Strikingly, the TNF responses that require endoso- 340:9. 2. Vassalli, P. 1992. The pathophysiology of tumor necrosis factor. Annu. Rev. mal acidification are identical with those that are inhibited Immunol. 10, 411. by D609. Therefore, we speculate that D609 may interfere with 3. Tracey, K. J., and A. Cerami. 1993. Tumor necrosis factor, other cytokines, and TNF signaling, at least in part, by inhibiting TNF-induced mac- disease. Annu. Rev. Cell Biol. 10:317. 4. Stout, R., and K. Bottomly. 1989. Antigen-specific activation of effector macro- ropinocytosis and thus preventing TNFR translocation to phages by interferon-␥-producing (Th1) T cell clones: failure of IL-4-producing endosomes. (Th2) T cell clones to activate effector functions in macrophages. J. Immunol. 142:760. TNF inhibits chemotaxis of phagocytes without affecting che- 5. Munoz-Fernandez, M. A., M. A. Fernandez, and M. Fresno. 1992. Synergism mokinesis (6–11, 39). To investigate the cytoskeletal basis of this between tumor necrosis factor-␣ and interferon-␥ on macrophage activation for the killing of intracellular Trypanosoma crusi through a nitric oxide-dependent inhibition, we tested the effects of TNF on CSF-1-induced actin mechanism. Eur. J. Immunol. 22:301. remodeling. We recently demonstrated that CSF-1 induces Cdc42- 6. Paulnock, D. M. 1992. Macrophage activation by T cells. Curr. Opin. Immunol. mediated filopodium formation and Rac-mediated lamellipodium 4:344. 7. Shalaby, M. R., M. A. Palladino, S.E. Hirabayshi, T. E. Eessalu, G. D. Lewis, formation in M␾ (28). In this study, we observed that TNF inhib- H. M. Shepard, and B. B. Aggarman. 1987. Receptor binding and activation of ited Cdc42- and CSF-1-induced filopodium production but did not polymorphonuclear neutrophils by tumor necrosis factor-␣. J. Leukocyte Biol. 41:196. impair Rac- or CSF-induced lamellipodium extension. Filopodia 8. Fast, D. J., P. M. Schlievert, and R. D. Nelson. 1988. Nonpurulent response to toxic shock syndrome toxin 1-producing Staphylococcus aureus. J. Immunol. have been suggested to play a pivotal role in chemotaxis as sensors Downloaded from for the chemotactic gradient; thus, the TNF-induced inhibition of 140:949. 9. Sullivan, G. W., J. Linden, E. L. Hewlett, H. T. Carper, J. B. Hylton, and Cdc42-mediated filopodium extension may explain the inhibitory G. L. Mandell. 1990. Adenosine and related compounds counteract tumor necro- effect of TNF on M␾ chemotaxis. The signal leading to an overall sis factor inhibition of neutrophil migration. J. Immunol. 145:1537. 10. Vollmer K. L., J. S. Alberts, H. T. Carper, and G. L. Mandell. 1992. Tumor necrosis decrease in F-actin in response to TNF may contribute to the in- factor-␣ decreases neutrophil chemotaxis to N-formyl-1-methionyl-1-leucyl-1-phe- ability of cells to produce filopodia, although it clearly does not nylalanine: analysis of single cell movement. J. Leukocyte Biol. 52:630. 11. Webb, S. E., J. W. Pollard, and G. E. Jones. 1996. Direct observation and quan-

prevent membrane ruffling. http://www.jimmunol.org/ tification of macrophage chemoattraction to the growth factor CSF-1. J. Cell Sci. The inhibitory effect of TNF on filopodium extension mapped to 109:793. the death domain of TNFR-1. Activation of p38 MAPK by TNF is 12. Tartaglia, L. A., R. F. Weber, I. S. Figari, C. Reynolds, M. A. Palladino, Jr., and D. V. Goedel. 1991. The two different receptors for tumor necrosis factor mediate similarly dependent upon this domain (40), and in the presence of distinct cellular responses. Proc. Natl. Acad. Sci. USA 88:9292. the p38 MAPK inhibitor SB203580, TNF no longer blocked filop- 13. Vandenabeele, P., W. Declercq, R. Beyaert, and W. Fiers. 1995. Two tumour odium production, suggesting that p38 MAPK mediates the inhi- necrosis factor receptors: structure and function. Trends Cell Biol. 5:392. 14. Adam-Klages, S., D. Adam, K. Wiegmann, S. Struve, W. Kolanus, bition of filopodium formation. In agreement with this, we ob- M. Schneider-Mergener, and M. Kro¨nke. 1996. FAN, a novel WD-repeat protein, served that cell-permeable ceramides, which stimulate p38 MAPK, couples the p55 TNF-receptor to neutral sphingomyelinase. Cell 86:937. 15. Hsu, H., J. Xiong, and D. V. Goedel. 1995. The TNF receptor 1-associated pro- impaired Cdc42-induced filopodium formation. Interestingly, both tein TRADD signals cell death and NF-␬B. Cell 81:495. actinomycin D and UV light strongly activate p38 MAPK in most 16. Rothe, M., M.-G. Pan, W.-J., Henzel, T. M. Ayres, and D. V. Goedel. 1994. The by guest on September 28, 2021 TNFR2-TRAF signaling complex contains two novel proteins related to baculo- cells, and have been reported to inhibit leukocyte chemotaxis (49, virus inhibitor or apoptosis proteins. Cell 83:1243. 50). This effect may also be due to p38 MAPK-mediated inhibition 17. Hsu, H., H.-B. Shu, M.-G., Pan, and D. V. Goedel. 1996. TRADD-TRAF2 and of filopodium formation. As the death domain of TNFR-1 medi- TRADD-FADD interactions define two distinct TNF receptor 1 signal transduc- tion pathways. Cell 84:299. ated both macropinocytosis and an inhibition of filopodium for- 18. Hsu, H., J. Huang, H.-B. Shu, M.-G., Pan, and D. V. Goedel. 1996. TNF-depen- mation, it is possible that macropinocytosis is causally associated dent recruitment of the protein kinase RIP to the TNF receptor-1 signalling com- plex. Immunity 4:387. with the lack of filopodia. However, SB203580 did not block mac- 19. Liu, Z.-G., H. Hsu, D. V. Goedel, and M. Karin. 1996. Dissection of TNF re- ropinocytosis, and therefore these two responses mediated by the ceptor 1 effector functions: JNK activation is not linked to apoptosis, while death domain are separable. Presumably, another target(s) of the NF-␬B activation prevents cell death. Cell 87:565. 20. Castellino, A. M., G. J. Parker, I. V. Boronenkov, R. A. Anderson, and death domain is involved in mediating the macropinocytotic re- M. V. Chao. 1997. A novel interaction between the juxtamembrane region of the sponse to TNF. p55 tumor necrosis factor receptor and phosphatidylinositol-4-phosphate 5-ki- nase. J. Biol. Chem. 272:5861. TNF is one of the most pleiotropic cytokines known, and its 21. Tartaglia L. A., T. M. Ayres, G. H. Wong, and D. V. Goedel. 1993. A novel effects on cell migration and locomotion vary considerably be- domain within the 55-kDa TNF receptor signals cell death. Cell 74:845. tween different cell types; for instance, TNF inhibits M␾ che- 22. Hall, A. 1994. Small GTP-binding proteins and the regulation of the actin cy- toskeleton. Annu. Rev. Cell Biol. 10:31. motaxis but acts as a chemoattractant for fibroblasts (51). Our 23. Tapon, N., and A. Hall. 1997. Rho, Rac, and Cdc42 GTPases regulate the orga- results show that distinct signals generated by different regions nization of the actin cytoskeleton. Curr. Opin. Cell. Biol. 9:86. ␾ 24. Kozma, R., S. Ahmed, A. Best, and L. Lim. 1995. The Ras-related protein of TNFR-1 have specific effects on actin organization in M Cdc42Hs and bradykinin promote formation of peripheral actin microspikes and (Fig. 6). The membrane proximal region of the receptor to- filopodia in Swiss 3T3 fibroblasts. Mol. Cell. Biol. 15:1942. gether with the death domain is required for the decrease in 25. Nobes, C. D., and A. Hall. 1995. Rho, Rac, and Cdc42 GTPases regulate the assembly of multimolecular focal adhesion complexes associated with actin stress F-actin, whereas the FAN-binding site is required for distinct fibers, lamellipodia, and filopodia. Cell 81:53. changes in actin organization involving a localized increase in 26. Ridley, A. J., and A. Hall. 1992. The small GTP-binding protein rho regulates the assembly of focal adhesions and actin stress fibers in response to growth factors. F-actin. The death domain alone is sufficient to mediate mac- Cell 70:389. ropinocytosis and an inhibition of filopodium formation. Hence, 27. Ridley, A. J., H. F. Paterson, C. L. Johnston, D. Diekman, and A. Hall. 1992. The small GTP-binding protein rac regulates growth factor-induced membrane ruf- different TNF-activated signal transduction pathways can exert fling. Cell 70:401. contrasting effects on the M␾ actin cytoskeleton, with the final 28. Allen, W. E., G. E. Jones, J. W. Pollard, and A. J. Ridley. 1997. Rho, Rac, and response presumably being dependent upon the relative abun- Cdc42 regulate actin organization and cell adhesion in macrophages. J. Cell Sci. 110:707. dance of secondary mediators and the strengths of these distinct 29. Lauffenburger, D. A., and A. F. Horwitz. 1996. Cell migration: a physically signals. Consequently, differential regulation of these signaling integrated molecular process. Cell 84:359. 30. Vandevoorde V., G. Haegeman, and W. Fiers. 1997. Induced expression of tri- pathways may explain the specific actions of TNF on cell lo- merized intracellular domains of the human tumor necrosis factor (TNF) p55 comotion in different cell types. receptor elicits TNF effects. J. Cell Biol. 137:1627. The Journal of Immunology 845

31. Brockhaus, M., H. J. Schoenfeld, E. J. Schlaeger, W. Hunziker, W. Lesslauer, and 42. Young, P., P. McDonnell, D. Dunnington, A. Hand, J. Laydon, and J. Lee. 1993. H. Loetscher. 1990. Identification of two types of tumor necrosis factor receptors Pyridinyl imidazoles inhibit IL-1 and TNF production at the protein level. Agents on human cell lines by monoclonal antibodies. Proc. Natl. Acad. Sci. USA 87: Actions 39:C67. 3127. 43. De Vos, K., V. Goossens, E. Boone, D. Vecammen, K. Vancompernolle, 32. Espevik, T., M. Brockhaus, H. Loetscher, U. Nonstad, and R. Shalaby. 1990. P. Vandenabeele, G. Haegeman, W. Fiers, and J. Grooten. 1998. The 55-kDa Characterization of binding and biological effects of monoclonal antibodies tumor necrosis factor receptor induces clustering of mitochondria through its against a human tumor necrosis factor receptor. J. Exp. Med. 171:415. membrane-proximal region. J. Biol. Chem. 273:9673. 33. Entwistle A., and M. Noble. 1994. Optimising the performance of confocal laser 44. Hartwig, J. H., G. M. Bokoch, C. L. Carpenter, P. A. Janmey, L. A. Taylor, scanning microscopes over the full field of view. J. Microsc. 175:238. A. Toker, and T. P. Stossel. 1995. Thrombin receptor ligation and activated Rac 34. Peppelenbosch, M. P., R.-G. Qiu, A. M. M. de Vries-Smits, L. G. J. Tertoolen, uncap actin filament barbed ends through phosphoinositide synthesis in perme- S. W. de Laat, F. McCormick, A. Hall, M. H. Symons, and J. L. Bos. 1995. Rac abilized human platelets. Cell 82:643. mediates growth factor-induced arachidonic acid release. Cell 81:849. 35. Wiegmann, K., S. Schu¨tze, T. Machleidt, D. Witte, and M. Kro¨nke. 1994. Func- 45. Rudel, T., and G. M. Bokoch. 1997. Membrane and morphological changes in tional dichotomy of neutral and acidic sphingomyelinases in tumor necrosis fac- apoptotic cells regulated by caspase-mediated activation of PAK2. Science 276: tor signalling. Cell 78:1005. 1571. 36. Watanabe-Fukunaga, R., C. I. Brannan, N. G. Copeland, N. A. Jenkins, and S. 46. Sells, M. A., U. G. Knaus, S. Bagrodia, D. M. Ambrose, G. M. Bokoch, and Nagata 1992. Lymphoproliferation disorder in mice explained by defects in Fas J. Chernoff. 1997. Human p21-activated kinase (Pak1) regulates actin organiza- antigen that mediates apoptosis. Nature 356:314. tion in mammalian cells. Curr. Biol. 7:202. 37. Swanson, J. A., and C. Watts. 1995. Macropinocytosis. Trends Cell Biol. 5:424. 47. Theriot, J. A. 1997. Accelerating on a treadmill: AOP/cofilin promotes rapid actin 38. Jaken, S. 1996. Protein kinase C isozymes and substrates. Curr. Opin. Cell Biol. filament turnover in dynamic cytoskeleton. J. Cell Biol. 136:1165. 8:168. 48. Ridley, A. J. 1994. Membrane ruffling and signal transduction. Bioessays 16:321. 39. Kharazami, A., H. Nielsen, and K. Brendtzen. 1988. Modulation of human neu- 49. Chang, F. Y., and M. F. Shaio. 1990. In vitro effect of actinomycin D on human trophil and monocyte chemotaxis and superoxide responses by recombinant TNF neutrophil function. Microbiol. Immunol. 34:311. and GM-CSF. Immunobiology 177:363. 40. Beyaert, R., A. Cuenda, W. Vanden Berghe, S. Plaisance, J. C. Lee, 50. Reibman, J., S. Meixler, T. C. Lee, L. I. Gold, B. N. Cronstein, K. A. Haines, G. Haegeman, P. Cohen, and W. Fiers. 1996. The p38/RK mitogen-activated S. L. Kolansinski, and G. Weismann. 1991. Transforming growth factor ␤1, a Downloaded from protein kinase pathway regulates interleukin-6 synthesis in response to tumour potent chemoattractant for human neutrophils, bypasses classic signal-transduc- necrosis factor. EMBO J. 15:1914. tion pathways. Proc. Natl. Acad. Sci. USA 88:6805. 41. Dudley, D. T., L. Pang, S. J. Decker, A. J. Bridges, and A. R. Saltiel. 1995. A 51. Postlethwaite, A. E., and J. M. Seyer. 1990. Stimulation of fibroblast chemotaxis synthetic inhibitor of the mitogen-activated protein kinase cascade. Proc. Natl. by human recombinant tumor necrosis factor and a synthetic TNF-31–68 peptide. Acad. Sci. USA 92:7686. J. Exp. Med. 172:1749. http://www.jimmunol.org/ by guest on September 28, 2021