Cells Effects on the Activation and Apoptosis of T Induces Opposing

Fibronectin-Associated Fas Ligand Rapidly Induces Opposing and Time-Dependent Effects on the Activation and Apoptosis of T Cells This information is current as of September 28, 2021. Alexandra Zanin-Zhorov, Rami Hershkoviz, Iris Hecht, Liora Cahalon and Ofer Lider J Immunol 2003; 171:5882-5889; ; doi: 10.4049/jimmunol.171.11.5882 http://www.jimmunol.org/content/171/11/5882 Downloaded from

References This article cites 40 articles, 17 of which you can access for free at:

http://www.jimmunol.org/content/171/11/5882.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 28, 2021

*average

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

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2003 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Fibronectin-Associated Fas Ligand Rapidly Induces Opposing and Time-Dependent Effects on the Activation and Apoptosis of T Cells1

Alexandra Zanin-Zhorov, Rami Hershkoviz, Iris Hecht, Liora Cahalon, and Ofer Lider2

Recently, it has been shown that Fas ligand (FasL) interacts with the extracellular matrix (ECM) protein fibronectin (FN), and that the bound FasL retains its cytotoxic efficacy. Herein, we examined the ramifications of FasL-ECM protein interactions throughout a specific time period, in the absence or presence of additional activating molecules, assuming that these complexed interactions occur during inflammation. We found that exposure of purified human T cells to FN-associated recombinant FasL for ␤ as brief as 5Ð10 min at 0.1Ð100 ng/ml induced their adhesion in 1 integrin- and FasR-dependent manners while activating the intracellular protein kinase, Pyk-2. The FN-associated FasL stops the CXCL12 (stromal cell-derived factor 1␣)-induced chemo- Downloaded from taxis of T cells by inhibiting the chemokine-induced extracellular signal-regulated kinase signaling and cytoskeletal rearrangement. This short term exposure of T cells to the FN-bound FasL (1 ng/ml), which was followed by T cell activation via the CD3 complex, resulted in 1) increased secretion of IFN-␥ (measured after 24 h), and 2) enhanced T cell apoptosis (measured after 72 h). Thus, in the context of inflamed ECM and depending on the time after FasL activation, its concentration, and the nature of other contextual mediators, FasL initially retains effector T cells at sites of inflammation and, later, induces T cell apoptosis and return

to homeostasis. The Journal of Immunology, 2003, 171: 5882Ð5889. http://www.jimmunol.org/

he functioning of immune cells, such as T cells, during coproteins of the ECM, and that such ECM-anchored FasL retains inflammation must be tightly controlled and time-limited. its cytotoxic potential (14). T These T cell functions depend on the existence of diverse We and other researchers have found that the ECM and, in par- inflammatory-modulating molecules, the cell’s state of activation, ticular, FN, can bind various growth factors, cytokines, and che- and the recognition of other collaborating cells and tissue compo- mokines and thereby affect immune cell activation and function in nents, including those of the extracellular matrix (ECM).3 One a site-restrictive manner (15Ð19). The ECM-anchored or associ- ␤ such mechanism that controls cell activation, function, and viabil- ated mediators affect monocyte and T cell activation, as well as 1 ity may be associated with the local interaction of T cells with Fas integrin functions associated with recognition of the surrounding by guest on September 28, 2021 ligand (FasL; CD95L). The engagement of stimulated T cells with tissue components, in addition to cell adhesion, migration, and cross-linked FasL leads to the activation of a caspase signaling cytokine and enzyme secretion (17, 20, 21). We have assumed that pathway that culminates in activation-induced apoptosis (1, 2). although the ECM-associated FasL can ultimately signal T cell However, in addition to its proapoptotic effects, it is now becoming death upon activation, in the short term it can also affect the proin- evident that FasL can function as an immune cell-activating factor flammatory functions of these cells. Based on our results, we pos- (3Ð6). Fas activation leads to T cell and neutrophil proliferation tulate that within the ECM, and depending on the time kinetics of (7), cytokine secretion (8, 9), angiogenesis and endothelial cell the interactions and the presence of other activatory moieties, FasL proliferation (9), and an increase in integrin expression on epithe- may differentially affect the activatory behavior of T cells and their lial cells (10). Although soluble FasL do not signal cell death (1, death and, consequently, the duration and intensity of the inflam- 2), it induces the chemotaxis of neutrophils (11, 12) and reduces matory reaction. neutrophil adhesion to endothelial cells (13). FasL can function in the context of the ECM; it has been found that FasL binds fi- Materials and Methods bronectin (FN) and vitronectin, two prototypic cell-adhesive gly- Reagents The following reagents and chemicals were purchased from the sources indicated: RPMI 1640 (Life Technologies, Paisley, U.K.); FCS, HEPES Department of Immunology, Weizmann Institute of Science, Rehovot, Israel buffer, antibiotics, and sodium pyruvate (Kibbutz Beit-Haemek, Israel); Received for publication July 2, 2003. Accepted for publication September 25, 2003. collagen type I (CO-I; Cellagen, ICN Pharmaceuticals, Costa Mesa, CA); The costs of publication of this article were defrayed in part by the payment of page collagen type VI (CO-IV; Sigma-Aldrich, Rehovot, Israel); FN (Chemicon, charges. This article must therefore be hereby marked advertisement in accordance Temecula, CA); laminin (LN); PMA, and the FN peptides, RGDS, RGES, with 18 U.S.C. Section 1734 solely to indicate this fact. and Leu-Asp-Val (LDV; Sigma-Aldrich). Recombinant human stromal ␣ ␣ 1 This work was supported by grants from the Minerva Foundation (Germany) and the cell-derived factor 1 (SDF-1 ), IL-2, soluble human extracellular FasL (a Israeli Science Foundation founded by the Israel Academy of Sciences and Human- monomeric 19.9-kDa protein, 175 aa residues), and TNF-␣ were purchased ities (Grant 1036/03). O.L. is the incumbent of the Weizmann League Career Devel- from PeproTech Asia (Rehovot, Israel). The ED50 of the cytotoxic effect of opment Chair in Children’s Diseases. this FasL is Ͻ10 ng/ml, corresponding to a sp. act. of Ͼ105 U/mg. mAbs ␤ 2 Address correspondence and reprint requests to Dr. Ofer Lider, Department of Im- directed against human 1 integrins (CD29, clone 3SS) and CD26 were munology, Weizmann Institute of Science, Rehovot 76100, Israel. E-mail address: obtained from Serotec (Oxford, U.K.). Neutralizing and activating mAb [email protected] against FasL and FasR, clones ZB4 and CH-11, respectively, were pur- 3 Abbreviations used in this paper: ECM, extracellular matrix; CO, collagen; ERK, chased from MBL (Nagoya, Japan). Capturing mAb anti-FasL (clone 4A5) extracellular signal-regulated kinase; FN, fibronectin; LDV, Leu-Asp-Val; LN, lami- and biotin-labeled mAb anti-FasL (clone 4H9) were purchased from MBL nin; PI3-K, phosphoinositol 3-kinase; PI-RGD, Arg-Gly-Asp; PKC, protein kinase C; (Nagoya, Japan). Polyclonal Ab against phosphorylated Pyk2 (clone SDF-1␣, stromal cell-derived factor 1␣. py881) was obtained from BioSource (Camarillo, CA), anti-total Pyk2

(clone N-19) was obtained from Santa Cruz Biotechnology (Santa Cruz, buffer was then added, and after boiling, the samples, containing equal CA), polyclonal Ab against phosphorylated extracellular signal-regulated amounts of proteins, were separated on 10% SDS-PAGE gel and trans- kinase2 (ERK2) was obtained from BioSource, anti-total ERK2 was ob- ferred to nitrocellulose membranes. The nitrocellulose membranes were tained from Sigma-Aldrich, and anti-human CXCR4 (clone 44708.111) blocked with TBST (20 mM Tris, 135 mM NaCl2, and 0.1 Tween 20, pH and anti-human CCR7 (clone 6B3) were purchased from R&D Systems 7.5) containing 5% low fat milk and probed with the following mAb in (Minneapolis, MN) and MBL (Nagoya, Japan), respectively. Inhibitors of blocking buffer: anti-phosphorylated Pyk2 (1.5 ␮g/ml), anti-total Pyk2 (0.2 intracellular signal transduction pathways, namely wortmanin and ␮g/ml), anti-phosphorylated ERK (0.2 ␮g/ml), or anti-total ERK (1/20,000 LY294002 (phosphoinositol 3-kinase (PI3-K) inhibitors), GF109203X from stock). Immunoreactive protein bands were visualized using a HRP- (protein kinase C (PKC) inhibitor), and pertussis toxin (G protein inhibitor) conjugated secondary Ab and the ECL system (20, 21). were obtained from Calbiochem (San Diego, CA). Actin polymerization Human T cells ϫ 6 o T cells (3 10 cells/ml) were preincubated with FasL (1 h, 37 C, 7% CO2 T cells were purified from healthy human donor peripheral blood. The humidified atmosphere) and treated with 500 ng/ml of SDF-1␣ for 15 s at o whole blood was incubated with RosetteSep human T cell enrichment 37 C and then fixed by the addition of a 3-fold volume of 3.7% PFA for o cocktail (StemCell Technologies, Vancouver, Canada) for 20 min at 22¡C. 10 min at 22 C. Next, the cells were extensively washed, and the mem- Then, the T cells were loaded on Lymphocyte Separation Medium (ICN branes were permeabilized; the T cells were stained with FITC-phalloidin Biomedicals, Irvine, CA), isolated by density centrifugation, and washed. (2 ␮g/ml), washed, and analyzed by FACScan at 525 nm, as we previously The purified cells obtained (Ͼ97% CD3ϩ T cells) were cultured in RPMI described (20). containing antibiotics and 10% heat-inactivated FCS. Cytokine secretion Binding of FasL to FN T cells (2 ϫ 106 cells/ml) were activated (10 min, 37oC) with the indicated

The flat-bottom microtiter well plates were precoated with FN, LN, CO-I, concentration of either soluble or FN-bound FasL in 24-well plates in a Downloaded from or CO-VI (10 ␮g/ml), and the remaining binding sites were blocked with serum-free medium containing 0.1% BSA. The cells were then washed and 0.1% BSA. rFasL, diluted in PBS containing 0.1% BSA, was incubated on replated in the same concentration on anti-CD3 mAb-precoated 24-well ␮ o the immobilized ECM protein (1 h at 37oC, 7% CO in a humidified at- plates (1 g/ml; non-tissue culture grade plates; 4 C, 24 h). The superna- 2 ␣ ␥ mosphere) and then gently washed. The remaining binding sites were tants were collected, and the cytokine content (TNF- and IFN- ) was blocked with 0.1% BSA. Analysis of recombinant human FasL binding to determined by ELISA, using the appropriate mAb, according to the man- FN was conducted as follows. Various concentrations of soluble FasL were ufacturer’s instructions. incubated (18 h, 4oC) on 96-well plates, which were previously coated with T cell apoptosis http://www.jimmunol.org/ capture mAb anti-human FasL (clone 4A5). Then, biotin-labeled anti- human FasL detection mAb (clone 4H9) was added to the washed wells, Human T cells (2 ϫ 106 cells/ml) were activated (10 min, 37oC) with and binding was analyzed using diluted HRP-labeled streptavidin and soluble, FN- or plastic-bound FasL in 24-well plates in RPMI medium 3,3Ј,5,5Ј-tetramethylbenzidine substrate (Chemicon, Temecula, CA). The containing 10% FCS. The cells were then washed and replated (2 ϫ 106 reaction was stopped with 50 ml of1MH2SO4, and the OD of the reaction cells/ml) on 24-well plates that were precoated with mAb anti-CD3 (1 was measured (at 450/570 nm) using a spectrophotometer. In parallel, var- ␮g/ml; non-tissue culture grade plates) and cultured for 72 h under tissue ious concentrations of soluble FasL were incubated (1 h, 37oC) on 96-well culture conditions. Then, the percentage of cells undergoing apoptosis was plates that were previously coated with FN (10 ␮g/ml, 1 h, 37oC). The determined using the annexin V detection assay. The cells were incubated binding to FN was measured using the detection Abs against FasL (clone (10 min in the dark) at room temperature in 200 ␮l of buffer containing 4A5) as described above. The amounts of FN-bound FasL were calculated FITC-conjugated human annexin V (5 ␮l; Bender MedSystem, San Bruno, using a binding curve of FasL to the mAb, clone 4A5. The percentage of CA). Finally, 10 ␮l of propidium iodide was added to each sample, and the by guest on September 28, 2021 FasL binding to FN was determined as FN-bound FasL/soluble FasL ϫ percentage of cells undergoing apoptosis was analyzed by FACScan at 525 100%. nm using CellQuest Software. The annexin Vϩ propidium iodideϪ cells corresponded to the apoptotic cells. The data are expressed as the percent- T cell adhesion and migration assays age of increase above untreated cells. Analysis of T cell adhesion to ECM components was determined as previously described (19Ð21). Briefly, 51Cr-labeled T cells were resuspended Results in RPMI medium supplemented with 1% HEPES buffer and 0.1% BSA FasL induces T cell adhesion to FN and LN (adhesion medium). Then, the indicated activators were added to microtiter First, we analyzed, using a specific mAb, the proportional quantity well plates (which were precoated with ECM proteins (10 ␮g/ml)) together with the cells (105 cells/well). The plates containing the test materials were of FasL binding to immobilized FN. The results, shown in Fig. 1A, o ϳ further incubated (37 C, 7% CO2 humidified atmosphere) and then gently indicate that 1% of the added protein actually bound FN. Hence, washed. The adherent T cells were lysed, removed, and counted using a in the experiments specified below, the amount of FN-bound FasL gamma counter (Packard, Downers Grove, IL). The results are expressed as was calculated according to the results of this molecular binding Ϯ the mean ( SD) percentage of bound T cells from quadruplicate wells. The assay. Next, we examined the possible proadhesive effects of sol- migration of 51Cr-labeled T cells was examined in a Transwell chemotaxis apparatus (6.5-mm diameter; Corning, Corning, NY), consisting of two uble and FN-bound FasL on human T cells. Briefly, T cells were compartments separated by polycarbonate filters (5-␮m pore size) pre- seeded onto FN-coated wells. Some wells were pretreated with treated (1 h, 37oC) with FN (25 ␮g/ml) (21, 22). 51Cr-labeled T cells (2 ϫ different amounts of FasL, which was washed away before cell 5 ␮ 10 in 100 l of adhesion medium) were added to the upper chambers with seeding (bound FasL); alternatively, the T cells were added to the or without activators. The bottom chambers contained 0.6 ml of the same medium, with or without SDF-1␣ (250 ng/ml). After incubation, the cells FN-coated wells together with soluble FasL. The amount of ad- that had transmigrated into the lower wells were collected, centrifuged, and herent T cells, measured after different time periods, is shown in lysed, and the radioactivity in the resulting supernatants was determined. Fig. 1, B and C. After a 40-min incubation, both soluble and FN- The percentage (ϮSD) of cell migration was calculated as the radioactivity bound FasL induced T cell adhesion to FN; the maximal adhesion counts in the lysates of the lower chambers (representative of the migrating of T cells to FN was achieved with 0.001Ð1 ng/ml of soluble FasL cells) divided by the total counts (representative of 2 ϫ 105 cells). and 0.1Ð10 ng/ml FN-bound FasL (Fig. 1B). Next, we measured Phosphorylation of Pyk-2 and ERK the ability of FasL to induce T cell adhesion to FN after a 10-min T cells were incubated in starvation medium (RPMI medium without se- exposure of the cells to the FN-FasL complex, either alone (soluble rum) for 48 h. Before testing, 5 ϫ 106 cells/sample were activated with of FasL) or with FN-bound FasL. The results demonstrate that such o FasL (37 C, 7% CO2 humidified atmosphere) with or without being fol- a short exposure of the T cells to both bound and soluble FasL lowed by incubation with SDF-1␣ (200 ng/ml, 10 min) or with immobi- induced appreciable levels of cell adhesion, a response that is man- lized mAb anti-CD3 (1 ␮g/ml, 30 min). Freezing the plates at Ϫ70oC for 10 min terminated the reaction. The thawed cells were incubated (60 min, ifested in a repeated pattern of bell-shaped dose-response curve; 4oC) in lysis buffer (21) and cleared by centrifugation (30 min, 14 ϫ 103 the maximal T cell adhesion induced by the mediator was also in rpm), and the supernatants were analyzed for protein content. Sample the range of 0.01 ng/ml (Fig. 1B). Note that a considerable number 5884 ECM-ASSOCIATED FasL AFFECTS T CELL FUNCTIONS

T cells remained viable (data not shown). Thus, similar to its intact soluble form, FN-anchored FasL can induce T cell adhesion. We then examined the ability of FasL to induce T cell adhesion to ECM glycoproteins other than FN. T cell adhesion was examined following a 10-min exposure of the cells to the soluble form of the mediator. We found that FasL induced an appreciable level of T cell adhesion to LN, but not to CO-I and CO-IV (Fig. 1D), whereas a T cell adhesion-inducing chemokine, SDF-1␣ (125 ng/ ml; 30-min pretreatment plus a 30-min adhesion assay), as well as PMA induced a signiﬁcant level of T cell adhesion to the four substrates (data not shown). Finally, we compared the time-dependent pattern of FasL-induced T cell adhesion to FN with the proadhesive effects of PMA and IL-2. IL-2 was chosen based on its ability to induce T cell adhesion to FN (22). The results, shown in Fig. 2, indicate that only FasL can induce T cell adhesion following as brief as a 5-min exposure of human T cells ( p Ͻ 0.05 compared with PMA and IL-2). After a 10-min exposure, the ability of FasL to induce cell

adhesion was equal to that of PMA, but signiﬁcantly superior to Downloaded from that of IL-2 ( p Ͻ 0.05). The three T cell activators reached their maximal proadhesive effect upon a 30-min incubation, an effect that gradually and moderately decreased with time, although FasL and PMA exerted a profound effect even after 180 min. Thus, in comparison with IL-2, FasL is capable of inducing rapid and pro-

longed T cell-FN interactions. http://www.jimmunol.org/

FasL-induced T cell adhesion to FN involves FasR recognition ␤ and FN-speciﬁc 1 integrins T cell recognition of and adhesion to FN are mediated primarily by ␣ ␤ ␣ ␤ the LDV- and Arg-Gly-Asp (RGD)-recognizing 4 1 and 5 1 integrins (CD29), respectively. To test whether these receptors are indeed involved in the apparent FasL-induced adhesion of T cells

to immobilized FN, we pretreated the responsive T cells with the by guest on September 28, 2021 indicated peptides and mAb as well as control molecules; their adhesion, induced by the soluble form of FasL, was measured 30 min after their seeding onto the FN-coated wells. The results, shown in Fig. 3A, demonstrate that the RGD and LDV peptides as well as mAb anti-CD29, but not the control peptide or mAb anti- CD26, inhibited T cell adhesion. Next we investigated whether the FasR can signal T cell adhesion. The T cells were treated with either the FasL inhibitory mAb

FIGURE 1. Induction by soluble or FN-associated recombinant human FasL of T cell adhesion to ECM component. A, Recombinant human FasL was added to immobilized FN for1hat37oC. FasL binding was then assessed by a pair of mAb against human FasL, as described in Materials and Methods. The percentage of binding was calculated as FN-bound FasL/ soluble FasL ϫ 100. BÐD, Purified human T cells were labeled with 51Cr, treated with FasL, and seeded onto FN-coated (B and C) or FN-, LN-, CO-I-, or CO-IV-coated (D) wells. Where indicated, some FN-coated wells were also pretreated with FasL, and the unbound mediator was removed after a short incubation (FN-bound FasL). The amount of adherent T cells was determined after 40 min (B)or10min(C and D). The mean Ϯ SD of five different experiments are shown. FIGURE 2. Soluble FasL induces rapid T cell adhesion to FN. Purified human T cells were labeled and treated with soluble FasL (1 ng/ml), PMA (50 ng/ml), or IL-2 (100 ng/ml) and seeded onto FN-coated microtiter wells, and the amount of adherent T cells was determined after different of T cells adhered to FN even if the cells were briefly (10 min) periods of incubation. Nonadherent T cells were removed by washing, and exposed to soluble FasL and washed before their seeding on the the number of remaining matrix-bound T cells was determined. The FN-coated wells, and that under these conditions the FasL-treated mean Ϯ SD of three different experiments are shown. The Journal of Immunology 5885

FasL-induced T cell adhesion involves PKC and PI3-K signaling, as well as Pyk-2 phosphorylation The adhesion of activated T cells to immobilized ECM glycoproteins requires speciﬁc intracellular signaling pathways that render ␤ nonresponding 1 integrins to their ECM recognition and binding states (23, 24). To verify that the FasL-induced T cell adhesion to FN involves cell activation and to discern which intracellular activation pathways are involved in this process, we exposed the T cells to FasL together with the inhibitory compounds. The results, showing that T cell adhesion was inhibited by GF109203X, wortmannin, and Ly294002, but not by pertussis toxin, indicate that the adhesive effects of FasL require the activation of PKC and PI3-K, but not G protein-coupled receptor signaling (Fig. 4A). The focal adhesion kinase, designated cytoplasmic tyrosine kinase 2 (Pyk-2), is phosphorylated upon PKC activation and links ␤ 1 integrins to multiple signaling pathways regulating adhesive and migratory processes in T cells (23, 24). Fig. 4, B and C, shows Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 3. FasL-induced T cell adhesion to FN is very late Ag 4- and 5-dependent and is mediated by Fas signaling pathway. A, Puriﬁed human T cells were radiolabeled, then pretreated with the indicated mAb or FN peptides. The cells were then treated with FasL (1 ng/ml) and seeded onto FN-coated microtiter wells, and the amount of cell adhesion was determined 30 min later. The mean Ϯ SD of three different experiments are shown. B, Puriﬁed human T cells were labeled and pretreated with the indicated mAb (10 ␮g/ml). The cells were treated with soluble FasL (1 ng/ml) or PMA (50 ng/ml) and seeded onto FN-coated wells. The amount of cell adhesion was determined 30 min later. C, The labeled T cells were treated with different concentrations of FasL or activatory anti-Fas Abs (CH-11) and seeded onto FN-coated wells, and the amount of cell adhesion was determined 30 min later. The mean Ϯ SD of four different experiments .p Ͻ 0.05 ,ء .are presented in B and C FIGURE 4. Signaling mechanisms involved in FasL-induced T cell adhesion to FN. A, T cells were pretreated with the indicated intracellular signal transduction inhibitors, wortmannin (5 nM), GF109203X (20 nM), pertussis toxin (2 ␮g/ml), and LY294002 (20 nM). The cells were then ZB4 (Fig. 3B) or the FasR activatory mAb CH-11 (Fig. 3C), and treated with FasL (1 ng/ml) and seeded onto FN-coated microtiter wells, their adhesion to FN was measured after a 30-min incubation. The and the amount of cell adhesion was determined 30 min later. The mean Ϯ results indicate that the CH-11 mAb indeed induced T cell adhe- SD of three different experiments are presented in A. T cells were exposed sion FasL, especially at 100-1000 ng/ml, and that mAb ZB4 in- to different concentrations of sFasL for 2 min (B)or10min(C) and then hibited FasL-induced, but not PMA-induced adhesion. The rela- lysed. The lysates were run on SDS-PAGE, transferred to nitrocellulose membranes, and immunoblotted with Ab anti-phospho-Pyk-2 (pPYK2) tively moderate proadhesive effect of mAb CH-11, compared with and anti-total Pyk-2 (tPYK2). The densitometric histograms of the exper- those of FasL, could be explained by its relatively high m.w. and iment are expressed as phosphorylated Pyk-2/total Pyk-2 ϫ 100%, and are complexed form, which probably affect its binding to CD95 and its from one experiment representative of three. These histograms are the av- ␣ ␤ ␣ ␤ ,ء .signaling. Thus, FasL exerts its 4 1 and 5 1 integrin-mediated erage calculated from the results of three different experiments Ϯ SD rapid proadhesive effect via the FasR. p Ͻ 0.05. 5886 ECM-ASSOCIATED FasL AFFECTS T CELL FUNCTIONS that brief treatments (10 and 2 min, respectively) of T cells with soluble FasL induced the phosphorylation of Pyk-2. In the 10-min exposure experiment, a concentration of Ͼ1 pg/ml FasL induced Pyk-2 phosphorylation, and in the 2-min exposure experiment, similar levels of phosphorylation occurred at Ͼ1 ng/ml. Note that in the 2-min assay, similar to T cell adhesion to FN, as shown in with Fig. 1C, Pyk-2 phosphorylation showed a bell-shaped dose- response curve with increasing amounts of soluble FasL. Thus, Pyk-2 phosphorylation accompanies the response of T cells to ef- fective concentrations of FasL.

FN-associated FasL stops SDF-1␣-induced T cell chemotaxis, actin cytoskeleton rearrangement, and signaling We then investigated whether FasL-FN interactions affect T cell chemotaxis induced by the inﬂammatory chemokine SDF-1␣ (25). To this end, we analyzed T cell migration in the Transwell apparatus through FN-coated membranes, in which net leukocyte migration toward a chemokine over time (3 h) was measured. First, we measured T cell chemotaxis induced by the FasL present in the Downloaded from lower chambers in different amounts and found that this mediator did not induce T cell migration, even after 3 h (not shown). Next, we measured T cell response to SDF-1␣ present in the lower wells in the absence or the presence of soluble or FN-bound FasL in the upper wells. We found that a substantial number of T cells mi-

grated when only SDF-1␣ was present in the lower chambers. http://www.jimmunol.org/ However, this migration of T cells was markedly reduced, in a dose-response manner, if either soluble or FN-bound FasL was present in the upper chambers. Maximal inhibition of T cell chemotaxis occurred with 100 ng/ml of soluble or bound FasL (Fig. 5A). A similar pattern of results were obtained when the chemokine ELC (CCL19) was used as the T cell chemoattractant (data not shown). These results were reminiscent of those previously reported by us where ECM-associated TNF-␣ was shown to stop the SDF-1␣- by guest on September 28, 2021 induced chemotaxis of CD4ϩ human T cells (26). To determine whether FN-associated FasL stops T cell chemotaxis similarly to TNF-␣, we compared the SDF-1␣-induced chemotaxis-modulating abilities of FasL to those of TNF-␣ or IL-2 (as a control proin- FIGURE 5. FasL pretreatment regulates T cell chemotaxis (A and B), flammatory and T cell adhesion-inducing cytokine) (22) present in actin polymerization (C), and ERK phosphorylation (D) induced by SDF- 1␣. A and B, T cells were labeled, washed, and added to the upper cham- the upper chambers. The results, shown in Fig. 5B, demonstrate ␣ bers of Transwell apparati with or without different concentrations of FasL that both FasL and TNF- , but not IL-2, induced a similar dose- (A and B), TNF-␣ (B), or IL-2 (B). The lower chambers contained the same dependent attenuated pattern of T cell chemotaxis; their maximal medium, with or without SDF-1␣ (250 ng/ml). After 3 h, the cells that had inhibitory effect occurred at 100 ng/ml ( p Ͻ 0.01). Thus, FN- transmigrated into the lower wells were collected, centrifuged, and lysed, associated FasL can stop T cell chemotaxis and, therefore, prob- and the radioactivity in the resulting supernatants was determined. The ably restricts the presence of T cells to defined ECM sites. mean Ϯ SD of five different experiments are shown. C, T cells were treated T cell adhesion to and migration through immobilized ECM (1 h) with the indicated concentrations of FasL, washed, and exposed for ligands require specific intracellular signaling and cellular polar- 15sat37¡C to SDF-1␣ (250 ng/ml). Next, the cells were fixed, and their ization, a process associated with the polymerization of the actin intracellular filamentous actin was stained with FITC-phalloidin and ana- cytoskeleton (27). To study the effect of FasL on SDF-1␣-induced lyzed by FACS. The results are expressed as the average mean fluores- cence Ϯ SD of three experiments. D, T cells were treated (1 h) with FasL, actin polymerization, T cells were left intact or were treated with ␣ ␣ washed, exposed for 10 min to SDF-1 (250 ng/ml), and lysed. The lysates sFasL (60 min) and then activated them by SDF-1 (100 ng/ml) were run on SDS-PAGE, transferred to nitrocellulose membranes, and im- for 15 s. Fig. 5C shows that soluble FasL at 10Ð100 ng/ml mark- munoblotted with Ab anti-phospho-ERK (pERK) and anti-ERK (tERK). edly inhibited actin polymerization induced by SDF-1␣. Next, we The SDS-PAGE analysis shown here is from one experiment representa- examined the effect of FasL on SDF-1␣-induced ERK phosphor- tive of three. The densitometric histograms of the experiment are expressed ylation. Activated ERK regulates cell adhesion and motility by as pERK/tERK ϫ 100%. These histograms are the average calculated from .p Ͻ 0.05 ,ء .enhancing the activity and phosphorylation of myosin and the po- the results of three different experiments Ϯ SD lymerization of actin fibers (28). The results, shown in Fig. 5D, indicate that FasL (1Ð100 ng/ml; 60 min) inhibits the SDF-1␣- migrating T cells by blocking their ability to react to the chemo- induced ERK phosphorylation in the T cells in a dose-dependent kine by undergoing polarization and intracellular ERK signaling. manner, reaching its maximal inhibitory effect at 100 ng/ml. When used alone, FasL did not induce ERK phosphorylation (not A brief exposure of T cells to FN-associated FasL and mAb anti- ␥ shown). Note that the results of the densitometric analysis, as de- CD3 results in enhanced IFN- secretion and T cell apoptosis picted in these figures, demonstrate the average of three different We analyzed the effect of a short exposure of T cells to FN-bound experiments. Thus, FasL delivers a stop signal to chemotactically FasL and compared it to the effect of an intact mediator. The T The Journal of Immunology 5887

Table I. FasL pretreatment up-regulates anti-CD3-induced T-cell secretion of IFN-␥ and apoptosisa

Treatment of T cells T Cell Responses following CD3 Stimulation

I. Cytokine secretion (ng/ml; 24 h) II. Apoptosis (% Substrate/form of FasL FasL (ng/ml) IFN-␥ TNF-␣ Increase (72 h))

None None 1.2 Ϯ 0.1 0.9 Ϯ 0.4 0 FN None 1.4 Ϯ 0.1 0.7 Ϯ 0.2 3 Ϯ 2 Soluble FasL 1 2.5 Ϯ 0.2a 0.8 Ϯ 0.1 49 Ϯ 6 FN-bound FasL 1 2.7 Ϯ 0.2a 0.8 Ϯ 0.1 43 Ϯ 4b

a T cells were seeded onto bare or FN-coated wells, treated (10-min) with the indicated concentrations of FN-bound FasL, collected, washed, and transferred to 24-well plates coated with mAb anti-CD3 (OKT3; 1 ␮g/ml) in serum-free medium (for cytokine secretion) or in medium containing 10% FCS (for apoptosis). The supernatants were collected after 24 hr and analyzed for IFN-␥ and TNF-␣ secretion. The percentage of cells undergoing apoptosis was determined, using the annexin V detection assay, after 72 h. The mean Ϯ SD of ﬁve different experiments is shown. a P Ͻ 0.05, comparing the activatory effects of FasL on IFN-␥ secretion. b P Ͻ 0.01, comparing the apoptotoic effect of FN-bound vs. intact FasL at 100 ng/ml.

cells were exposed to FasL for 10 min in its intact or FN-bound appears to be an equal activator of T cells to the soluble mediator. Downloaded from form and then collected, washed, and activated with mAb anti- Note that 10-min exposure of T cells to intact FN did not result in CD3. T cell secretion of IFN-␥ and TNF-␣ was analyzed 24 h IFN-␥ secretion or apoptosis. Hence, a brief exposure of human T later, and T cell apoptosis was measured after 72 h. The results cells to FN-associated FasL and a subsequent incubation period of indicate that FasL, either intact or FN-bound, induced the secretion 1Ð3 days with CD3-activating mAb definitively affects their proin- of IFN-␥, but not TNF-␣, with maximal and almost equal effects flammatory capabilities (e.g., IFN-␥ secretion) and later their cell at 1 and 100 ng/ml for intact, soluble FasL or 1 ng/ml for the death, respectively. http://www.jimmunol.org/ FN-bound mediator. Next, we measured the effect of a brief exposure of T cells to Discussion FasL on T cell apoptosis induced by CD3 activation. The results In this study we have shown that as brief as a Ͼ10-min exposure indicate that FN-bound FasL indeed induced cell death. The max- of human T cells to FN (or LN)-associated recombinant FasL in- imal apoptotic effect occurred at 1Ð100 ng/ml for intact FasL and duces T cell activation and potentiates their proinflammatory ca- 1 ng/ml for FN-complexed FasL (Table I). Thus, FN-bound FasL pacities, such as adhesion to the ECM ligands, IFN-␥ secretion, by guest on September 28, 2021

FIGURE 6. Schematic model of the time dependency of FasL-induced T cell activation and apoptosis at sites of inflammation. 5888 ECM-ASSOCIATED FasL AFFECTS T CELL FUNCTIONS retention in inflamed areas, and, later, T cell apoptosis (Fig. 6). In ously shown to desensitize chemokine receptor signaling (39, 40). these reactions, FasL may affect T cell behavior when used alone The FasL-FN-induced retention of T cells within the inflamed spot on FN (adhesion), when the FN-associated FasL is presented to the probably enables their proinflammatory functioning, as it leads to cells before their exposure to mAb anti-CD3 (cytokine secretion their activation and the secretion of IFN-␥. and apoptosis), or when the T cells are first exposed to the matrix- Several mediators, including TNF-␣, TGF-␤, and macrophage associated mediator and then to a chemokine (chemotaxis). inflammatory protein-1␤, can bind various components of the The environmental conditions affect the mechanisms by which ECM (16Ð18). The ECM could thus serve as a transitory storage FasL confers immune privilege (tolerance) or induces inflamma- site for these inflammatory mediators, retaining them close to ef- tion (4, 5). For example, it has been shown that the exact site of fector cells, where they ensure cellular responses even with minute inoculation of colon carcinoma cells determines their fate (6). It amounts of the cytokines, because they are presented in a rather has also been reported that 1) depending on its form, e.g., mem- concentrated form to the attached leukocytes. These ECM-cyto- brane-expressed or soluble, FasL induces opposing effects on in- kine complexes may also restrict or modulate the cytokines’ ac- flammation and tumor cell survival (9, 29); and 2) the opposing tivities at target sites and differentially regulate the distribution, effects of FasL depend on the subpopulation of responding T cells availability, and activities of the matrix-associated mediators. (30). Also, the inflammatory or death-inducing capacities of FasL FasL, which may be enzymatically removed from tissue-invading depend on the presence of other inflammatory-regulating cyto- leukocytes and implanted into its ECM microenvironment (14), kines; TGF␤ synergizes with FasL in inducing tolerance (6). We can therefore behave like one of these inflammatory-modulating postulate that the proinflammatory or apoptotic effects of FasL cytokines. In their study on concentrated FasL-containing super- depend on its time of exposure, form (e.g., soluble or FN-associ- natants of CT26.CD95L cells, Aoki and co-workers found that the Downloaded from ated), amount, copresence of migration-inducing chemokines, FN-bound FasL undergoes oligomerization, which probably ac- TCR-activating agents, and sequence of analysis of T cell behavior counts for its enhanced activity, even though it is present on the with time following recognition and ligation of the FasL-ECM ECM protein and is active in low amounts. complex. Our findings indicate that the diverse effects of the ECM-asso- Our results indicate that when used alone, the ability of FasL to ciated FasL should be differentially examined on a time scale of

induce T cell adhesion to FN (or LN) involves the activation of minutes, hours, and days (Fig. 6). We suggest that the two seem- http://www.jimmunol.org/ ␤ specific 1 integrins. This activation, which was blocked by spe- ingly opposing effects of FasL, i.e., induction of proinflammatory cific inhibitors of PI3-K and PKC, was also associated with the behavior (life) and apoptosis (death), which may occur in the con- phosphorylation of Pyk-2, a member of the focal adhesion kinase text of the ECM in signal-rich extravascular sites of inflammation, family of molecules. Pyk-2 is phosphorylated upon PKC activa- should be regarded as a rather sophisticated mechanism by which ␤ tion; it links 1 integrins to multiple signaling pathways that reg- the very same molecule that induces inflammation can later enable ulate T cell adhesion and migration. Such intracellular processes healing and ensure the return of tissue homeostasis. are also involved in the regulation of mitogen-activated protein kinases, such as ERK-2, and Jun-NH kinase pathways (28, 31, 2 References by guest on September 28, 2021 32). ERK regulates immune cell adhesion and motility by enhanc- 1. Maher, S., D. Toomey, C. Condron, and D. Bouchier-Hayes. 2002. Activation- ing the activity of the myosin light chain kinase, myosin phos- induced cell death: the controversial role of Fas and Fas ligand in immune priv- phorylation, and consequent polarization of the actin fibers (27, 28, ilege and tumour counterattack. Immunol. Cell. Biol. 80:131. 2. Green, D. R., and T. A. Ferguson. 2001. The role of Fas ligand in immune 33). Our results also indicate that when the migrating T cells re- privilege. Nat. Rev. Mol. Cell. Biol. 2:917. sponded to both SDF-1␣ and FasL, the latter inhibited the chemo- 3. Alderson, M. R., R. J. Armitage, E. Maraskovsky, T. W. Tough, E. Roux, kines’ ability to induce ERK signaling and the rearrangement of K. Schooley, F. Ramsdell, and D. H. Lynch. 1993. Fas transduces activation signals in normal human T lymphocytes. J. Exp. Med. 178:2231. the actin cytoskeleton. 4. Chen, J-J., Y. Sun, and G. J. Nabel. 1998. Regulation of the proinflammatory A shown here, a short exposure of T cells to FN-associated effects of Fas ligand (CD95L). Science 282:1714. FasL, followed by activation via the CD3 pathway, was sufficient 5. Hohlbaum, A. M., R. R. Saff, and A. Marshak-Rothstein. 2002. Fas ligand: iron ␥ fist or Achilles’ heel. Clin. Immunol. 103:1. to induce the secretion of IFN- (24 h) and later (72 h) resulted in 6. Budd, R. C. 2002. Death receptors couple to both cell proliferation and apoptosis. cell death. It is reasonable that these effects of FN-associated FasL J. Clin. Invest. 109:437. involve NF-␬B and ERK signaling (34Ð36). However, in contrast 7. Kennedy, N. J., T. Kataoka, J. J. Tscjopp, and R. C. Budd. 1999. Caspase acti- ␥ vation is required for T-cell proliferation. J. Exp. Med. 190:1891. to its effect on IFN- secretion, a short exposure of T cells to FasL 8. Hohlbaum, A. M., M. S. Gregory, S. T. Ju, and A. Marshak-Rothstein. 2001. Fas did not affect the secretion of TNF-␣ regardless of whether FasL ligand engagement of resident peritoneal macrophages in vivo induces apoptosis was presented to the cells in the presence or the absence of FN. and the production of neutrophil chemotactic factors. J. Immunol. 167:6217. 9. Hohlbaum, A. M., S. Moe, and A. Marshak-Rothstein. 2000. Opposing effects of The reason for these diverse effects of FasL (and FN-specific in- transmembrane and soluble Fas ligand expression in inflammation and tumor cell tegrins) on cytokine secretion is presently not clear. However, it survival. J. Exp. Med. 191:1209. has been documented that the ligation of TNF-␣ by TNF-R1 leads 10. Jarad, B., S. Khan, J. deVore, H. Miao, K. Wu, S. L. Nishimura, B. A. Wible, M. Konieczkowski, J. R. Sedor, and J. R. Schelling. 2002. Fas activation induces to protection against T cell death (37). We postulate that such an renal tubular epithelial cell b8 integrin expression and function in the absence of antiapoptotic, downstream effect of FasL, presented to the T cells apoptosis. J. Biol. Chem. 49:47826. 11. Biancone, L., A. De Martino, V. Orlandi, P. G. Conaldi, A. Toniolo, and within an inflamed site, is not desirable; T cell apoptosis only G. Camussi. 1997. Development of inflammatory angiogenesis by local stimu- occurs following the prior functioning of T cells as modulators of lation of Fas in vivo. J. Exp. Med. 186:157. inflammation. 12. Ottonello, L., G. Tortolina, M. Ameloti, and F. Dallegri. 1999. Soluble Fas ligand is chemotactic for human neutrophilic polymorphonuclear leukocytes. J. Immu- An additional role of FasL, secreted and integrated into the nol. 162:3601. ECM at sites of inflammation, may be to arrest the chemokine- 13. Greenstein, S., J. Bernard, K. Zhou, M. Fong, and B. Hendey. 2000. Fas activa- induced migration of T cells at these sites, where their effector tion reduces neutrophil adhesion to endothelial cells. J. Leukocyte Biol. 68:715. 14. Aoki, K., M. Kurooka, J.-J. Chen, J. Petryniak, E. G. Nabel, and G. J. Nabel. functions are mostly required. It has been shown that such an an- 2001. Extracellular matrix interacts with soluble CD95L: regulation and enhance- chorage of T cells may also occur after ligation of TCR or TNF ment of cytotoxicity. Nat. Immunol. 2:333. receptor (26, 38). As shown here, FasL-induced T cell adhesion to 15. Li, J., Y. P. Zhang, and R. S. Kirsner. 2003. Angiogenesis in wound repair: angiogenic growth factors and the extracellular matrix. Microsc. Res. Technol. FN also involves PKC activation. This may explain the ability of 60:107. FasL to inhibit T cell chemotaxis, as PKC activators were previ- 16. Nathan, C., and M. Sporn. 1991. Cytokines in context. J. Cell Biol. 113:981. The Journal of Immunology 5889

17. Savino, W., D. A. Mendes-da-Cruz, J. S. Silva, M. Dardenne, and human Fas ligand is responsible for its inflammatory activity. Eur. J. Immunol. V. Cotta-de-Almeida. 2002. Intrathymic T-cell migration: a combinatorial inter- 31:2504. play of extracellular matrix and chemokines? Trends Immunol. 23:305. 30. Desbarats, J., T. Wade, W. F. Wade, and M. K. Newell. 1999. Dichotomy be- 18. Vaday, G. G., S. Franitza, H. Schor, I. Hecht, A. Brill, L. Cahalon, R. Hershkoviz, tween naive and memory CD4ϩ T-cell responses to Fas engagement. Proc. Natl. and O. Lider. 2001. Combinatorial signals by inflammatory cytokines and che- Acad. Sci. USA 15:5167. mokines mediate leukocyte interactions with extracellular matrix and secretion of 31. Van Seventer, G. A., M. M. Mullen, and L. M. Van Seventer. 1998. Pyk2 is enzymes. J. Leukocyte Biol. 69:885. differentially regulated by integrin- and CD28-mediated co-stimulation in human 19. Pelletier, A. J., L. J. van der Laan, P. Hildbrand, M. A. Siani, D. A. Thompson, CD4 T lymphocytes. Eur. J. Immunol. 28:3967. P. E. Dawson, B. E. Torbett, and D. R. Salomon. 2000. Presentation of chemo- 32. Seig, D. J., C. R. Hauk, D. Ilic, C. K. Klingbell, E. Schaefer, C. H. Damsky, and kine SDF-1␣ by fibronectin mediates directed migration of T cells. Blood D. D. Schlaepfer. 2000. FAK integrates growth-factor and integrin signals to 96:2682. promote cell migration. Nat. Cell. Biol. 2:249. 20. Franitza, S., O. Kollet, A. Brill, G. G. Vaday, I. Petit, T. Lapidot, R. Alon, and O. Lider. 2002. TGF-␤1 enhances SDF-1␣-induced chemotaxis and homing of 33. Samstag, Y., S. M. Eibert, M. Klemke, and G. H. Wabnitz. 2003. Actin cytoskel- naõ¬ve T cells by up-regulating CXCR4 expression and downstream cytoskeletal etal dynamics in T lymphocyte activation and migration. J. Leukocyte Biol. effector molecules. Eur. J. Immunol. 32:193. 73:30. 21. Hecht, I., L. Cahalon, R. Hershkoviz, A. Lahat, S. Franitza, and O. Lider. 2003. 34. Kataoka, T., R. C. Budd, N. Holler, M. Thome, F. Martinon, M. Irmler, K. Burns, Heterologous desensitization of T-cell functions by CCR5 and CXCR4 ligands: M. Hahne, N. Kennedy, M. Kovcsovics, et al. 2000. The caspase-8 inhibitor FLIP inhibition of cellular signaling, adhesion, and chemotaxis. Int. Immunol. 15:29. promotes activation of NF-␬B and ERK signaling pathways. Curr. Biol. 10:640. 22. Ariel, A., E. J. Yavin, R. Hershkoviz, A. Avron, S. Franitza, I. Hardan, 35. Soderstrom, T. S., M. Poukkula, T. H. Holmstrom, K. M. Heiskanen, and L. Cahalon, M. Fridkin, and O. Lider. 1998. IL-2 induces T-cell adherence to J. E. Erikson. 2002. Mitogen-activated protein kinase/ERK signaling in activated extracellular matrix: inhibition of adherence and migration by IL-2 peptides gen- T cells abrogates TRAIL-induced apoptosis upstream of the mitochondrial am- erated by leukocyte elastase. J. Immunol. 161:2465. plification loop and caspase-8. J. Immunol. 169:2851. 23. Damsky, C. H., and D. Ilic. 2002. Integrin signaling: it’s where the action is. 36. Xiao, S., S.-S. Jodo, J. Su, A. Marshak-Rothstein, and S.-T. Ju. 2002. A novel Curr. Opin. Cell. Biol. 14:594. signaling mechanism for soluble CD95 ligand: synergy with anti-CD95 mono- ␤ 24. Woods, M. L., and Y. Shimizu. 2001. Signaling networks regulating 1 integrin- clonal antibodies for apoptosis and NF-␬B nuclear translocation. J. Biol. Chem. Downloaded from mediated adhesion of T lymphocytes to extracellular matrix. J. Leukocyte Biol. 277:50907. 69:874. 37. Natoli, G., A. Costanzo, F. Guido, F. Moretti, and M. Levero. 1998. Apoptotic, 25. Nanki, T., K. Hayashid, H. S. El-Gabalawy, S. Suson, K. Shi, H. J. Girschick, non-apoptotic, and anti-apoptotic pathways of tumor necrosis factor signaling. S. Yavuz, and P. E. Lipsky, P. E. 2000. Stromal cell-derived factor-1-CXC che- Biochem. Pharmacol. 56:915. mokine receptor 4 interactions play a central role in CD4ϩ T-cell accumulation in rheumatoid arthritis synovium. J. Immunol. 165:6590. 38. Dustin, M. L., S. K. Bromley, Z. Kan, D. A. Peterson, and E. R. Unanue. 1997. 26. Franitza, S., R. Hershkoviz, N. Kam, N. Lichtenstein, G. G. Vaday, R. Alon, and Antigen receptor engagement delivers a stoppage signal to migrating T lympho- O. Lider. 2000. TNF␣ associated with extracellular matrix fibronectin provides a cytes. Proc. Natl. Acad. Sci. USA 94:3909. 39. Didsbury, J. R., R. J. Uhing, E. Tomhave, C. Gerard, N. Gerard, and stop signal for chemotactically migrating T cells. J. Immunol. 165:2738. http://www.jimmunol.org/ 27. Serrador, J. M., M. Nieto, and F. Sanchez-Madrid. 1999. Cytoskeletal rearrange- R. Snyderman. 1991. Receptor class desensitization of leukocyte chemoattractant ment during migration and activation of T lymphocytes. Trends Cel. Biol. 9:228. receptors. Proc. Natl. Acad. Sci. USA 88:11564. 28. Sanchez-Madrid, F., A. Hidalgo, and J. Teixido. 1999. Leukocyte polarization in 40. Wang, J.-F., I.-W. Park, and J. E. Groopman. 2000. Stromal cell-derived fac- cell migration and immune reactions. EMBO J. 18:501. tor-1␣ stimulates tyrosine phosphorylation of multiple focal adhesion proteins 29. Shudo, K., K. Kinoshita, R. Imamura, H. Fan, K. Hasumoto, M. Tanaka, and induces migration of hematopoietic progenitor cells: roles of phosphoinosi- S. Nagata, and T. Suda. 2001. The membrane-bound but not the soluble form of tide-3 kinase and protein kinase C. Blood 95:2505. by guest on September 28, 2021