β1 Integrin Cross-Linking Inhibits CD16-Induced Phospholipase D and Secretory Phospholipase A 2 Activity and Granule Exocytosis in Human NK cells: Role of This information is current as Phospholipase D in CD16-Triggered of September 25, 2021. Michele Milella, Angela Gismondi, Paola Roncaioli, Gabriella Palmieri, Stefania Morrone, Mario Piccoli, Luigi

Frati, Maria Grazia Cifone and Angela Santoni Downloaded from J Immunol 1999; 162:2064-2072; ; http://www.jimmunol.org/content/162/4/2064 http://www.jimmunol.org/ References This article cites 45 articles, 29 of which you can access for free at: http://www.jimmunol.org/content/162/4/2064.full#ref-list-1

Why The JI? Submit online.

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

by guest on September 25, 2021 • No Triage! Every submission reviewed by practicing scientists

• 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. ␤ 1 Integrin Cross-Linking Inhibits CD16-Induced Phospholipase D and Secretory Phospholipase A2 Activity and Granule Exocytosis in Human NK cells: Role of Phospholipase D in CD16-Triggered Degranulation1

Michele Milella,* Angela Gismondi,* Paola Roncaioli,† Gabriella Palmieri,*‡ Stefania Morrone,* Mario Piccoli,* Luigi Frati,*§ Maria Grazia Cifone,† and Angela Santoni2*

Recent data indicate that integrin-generated signals can modulate different -stimulated cell functions in both a positive ␤ ␣ ␤ ␣ ␤ (costimulation) and a negative (inhibition) fashion. Here we investigated the ability of 1 integrins, namely 4 1 and 5 1

␤ Downloaded from fibronectin receptors, to modulate CD16-triggered phospholipase activation in human NK cells. 1 integrin simultaneous cross- linking selectively inhibited CD16-induced phospholipase D (PLD) activation, without affecting either phosphatidylinositol-phos- pholipase C or cytosolic phospholipase A2 (PLA2) enzymatic activity. CD16-induced secretory PLA2 (sPLA2) release as ␤ well as its enzymatic activity in both cell-associated and soluble forms were also found to be inhibited upon 1 integrin coen- gagement. The similar effects exerted by specific PLD pharmacological inhibitors (2,3-diphosphoglycerate, ethanol) suggest that in our experimental system, sPLA2 secretion and activation are under the control of a PLD-dependent pathway. By using phar- macological inhibitors (2,3-diphosphoglycerate, wortmannin, ethanol) we also demonstrated that PLD activation is an important http://www.jimmunol.org/ step in the CD16-triggered signaling cascade that leads to NK cytotoxic granule exocytosis. Consistent with these findings, fi- bronectin receptor engagement, by either mAbs or natural ligands, resulted in a selective inhibition of CD16-triggered, but not of PMA/ionomycin-induced, degranulation that was reversed by the exogenous addition of purified PLD from Streptomyces chromofuscus. The Journal of Immunology, 1999, 162: 2064–2072.

ntegrins are a superfamily of ␣␤ heterodimeric adhesion re- different cell functions, such as migration, adhesion, proliferation, ceptors involved in both cell-cell and cell-extracellular ma- differentiation, apoptosis, and specific gene expression. In both trix (ECM)3 interactions. Members belonging to the same outside/in and inside/out signaling, integrins exhibit a functional I by guest on September 25, 2021 family share a common ␤ subunit paired with different ␣-chains cross-talk with diverse activation molecules, such as growth factor and are endowed with distinct, although overlapping, ligand bind- or Ag receptors (1, 2). ing specificities. Cell differentiation and activation tightly regulate Integrin-mediated costimulation of leukocyte activities has been integrin expression and avidity for their ligands. On the other hand, widely studied at both molecular and functional levels (3). Increas- integrin engagement by natural ligands or specific mAbs trans- ing evidence indicates that integrins are also capable of transduc- duces intracellular signals that are important for regulating many ing signals that negatively affect several receptor-stimulated cell functions, such as proliferation, degranulation, cytokine produc- tion, and specific gene expression, in different cell types (4–15). *Department of Experimental Medicine and Pathology, Istituto Pasteur-Fondazione Cenci Bolognetti, University of Rome “La Sapienza,” Rome, Italy; †Department of However, both the biochemical nature of integrin-elicited negative Experimental Medicine, University of L’Aquila, L’Aquila, Italy; ‡Biotechnology Sec- signals and the molecular levels at which they interfere with § tion, Institute for the Study and Cure of Tumors, Genoa, Italy; and Mediterranean known activator pathways are poorly defined. Lipid signaling has Institute of Neuroscience “Neuromed,” Pozzilli, Italy been recently indicated as one possible point of interference be- Received for publication August 7, 1998. Accepted for publication November 9, 1998. tween integrin and other receptor-initiated pathways. Inhibition of ␤ The costs of publication of this article were defrayed in part by the payment of page proliferation induced by the anti- 1 integrin mAb K20 is charges. This article must therefore be hereby marked advertisement in accordance indeed accompanied by cAMP accumulation as well as by a de- with 18 U.S.C. Section 1734 solely to indicate this fact. crease in the production of phosphatidic acid (PA) and diacylglyc- 1 This work was supported by grants from the Italian Association for Cancer Research erol (DAG) (4, 5), pointing to an interference with phospholipid and the Ministero dell’Universita`e della Ricerca Scientifica e Tecnologica (40% from MURST and 60% from Facolta`and Ateneo). turnover; in addition, cytoskeletal translocation of phosphatidyl- 2 Address correspondence and reprint requests to Dr. Angela Santoni, Department of inositol 3-kinase (PI 3-kinase) is up-regulated upon disruption of Experimental Medicine and Pathology, University of Rome “La Sapienza,” Viale GPIIb/IIIa-mediated aggregation in von Willebrand factor-stimu- Regina Elena 324, 00161 Rome, Italy. E-mail address: [email protected] lated platelets (16). 3 Abbreviations used in this paper: ECM, extracellular matrix; PA, phosphatidic acid; NK cells are a discrete CD3ϪCD16ϩCD56ϩ subset DAG, diacylglycerol; PI 3-kinase, phosphatidylinositol 3-kinase; PLC, phospholipase C; PLD, phospholipase D; PLA2, phospholipase A2; PYK-2, proline-rich tyrosine capable of lysing a broad range of tumor, virus-infected, and im- kinase-2; cPLA2, cytosolic phospholipase A2; sPLA2, secretory phospholipase A2; mature hemopoietic cells without prior sensitization in a non- GAM, F(abЈ) of goat anti-mouse Ig; pBPB, p-bromophenacyl bromide; DPG, 2,3- 2 MHC-restricted or Ab-dependent fashion (Ab-dependent cellular diphosphoglycerate; PC, phosphatidylcholine; PtdIns(4,5)P2, phosphatidylinositol bisphosphate; PtdIns(3,4,5)P3, phosphatidylinositol trisphosphate; Ins(1,4,5)P3, ino- cytotoxicity). Functional responses, such as exocytosis of cyto- sitol trisphosphate; LT, leukotriene; PAF, platelet-activating factor; PEt, phosphati- toxic granule content and cytokine production, are the final result dylethanol; AA, arachidonic acid; FN, fibronectin; BLT, N-␣-benzyloxycarbonyl-L- lysine thiobenzyl ester. of NK cell activation elicited by different stimuli (17, 18). CD16,

Copyright © 1999 by The American Association of Immunologists 0022-1767/99/$02.00 The Journal of Immunology 2065 the low affinity receptor for the Fc fragment of IgG (Fc␥RIIIA), is PLD assay a major signaling structure on NK cells capable of triggering mul- Human NK cells (500 ϫ 106) were preincubated with 1 mCi of [3H]oleic tiple biochemical pathways, among which the activation of several acid in serum-free medium for2hat37°C and then washed three times enzymes involved in phospholipid turnover, including phosphati- with medium containing 0.01% BSA. Radiolabeled cells (10 ϫ 106) were dylinositol-phospholipase C (PI-PLC) (19, 20), PLD (21), and both incubated with saturating concentrations of anti-CD16 (B73.1), anti-CD56 secretory and cytosolic PLA (22). Moreover, phospholipase acti- (B159.5.2), anti-MHC class I (W6/32), or anti-integrin (4B4, TS2/16, 2 HP2/1, SAM-1) mAbs for 15 min at 37°C in RPMI 1640 medium, then vation and lipid second messenger production are thought to be washed and incubated with GAM (1 ␮g/106 cells) for 5 min at 37°C in involved in the regulation of several NK cell functions (19–23). medium containing 0.5% ethanol. Stimulation was stopped by the addition ␤ Human NK cells express several members of the 1 family of of 2 vol of ice-cold methanol. In some experiments cells were treated with integrins, which have been suggested to play a major role in the pBPB (10–20 ␮M), DPG (2.5–10 mM), or wortmannin (0.3–300 nM) for ␤ 15 min at 37°C before stimulation. Lipids were extracted as described by regulation of their migration and functions (24–27). 1 integrin Bligh and Dyer (32) and were separated by TLC with a mobile phase ligation on human NK cells transduces different biochemical sig- system consisting of chloroform/pyridine/formic acid (50/30/7). Standards nals, including calcium mobilization, tyrosine phosphorylation of were chromatographed in parallel. TLC plates were then autoradiographed, various substrates, among which is proline-rich tyrosine kinase-2 and the phospholipid bands of interest were scraped and counted in a beta (PYK-2), and activation of the Ras/mitogen-activated protein ki- counter. PLD activity was quantitated by measuring the production of ␥ phosphatidylethanol (PEt) and was expressed as the percent increase above nase pathway, which leads to IFN- production (28–31). More- the basal level of untreated samples. ␤ over, 1 integrin cross-linking by either mAbs or natural ligands costimulates NK cytotoxic functions (28).

␤ PI-PLC and PLA2 assays Downloaded from Here we provide evidence that 1 integrin coengagement spe- cifically interferes with CD16-elicited PLD activation, without af- Human NK cells (10 ϫ 106) were incubated with saturating concentrations fecting PI-PLC and cPLA2 activity. Both sPLA2 activity and gran- of anti-CD16 (B73.1), anti-CD56 (B159.5.2), anti-MHC class I (W6/32), ule exocytosis are also inhibited as a consequence of PLD or anti-integrin (4B4, TS2/16, HP2/1, SAM-1) mAbs for 15 min at 37°C in RPMI 1640 medium. Cells were washed and incubated with GAM (1 ␮g/ inhibition, suggesting a role for this enzyme in CD16-triggered NK 106 cells) for different time periods at 37°C in RPMI 1640 medium. Stim- cell degranulation. ulation was stopped by the addition of ice-cold medium and centrifugation

at 500 ϫ g for 1 min at 4°C. In some experiments, cells were treated with http://www.jimmunol.org/ pBPB (10–20 ␮M) or DPG (2.5–10 mM) for 15 min at 37°C before stim-

Materials and Methods ulation. Supernatants were collected and assayed for the presence of sPLA2 Abs and reagents activity; sPLA2 protein release was assessed by ELISA, and the production of peptido-LT, LTB4, and PAF was tested by RIA. Cell pellets were re- The following mouse mAbs were used: anti-CD3 (Leu 4), anti-CD16 (Leu suspended in 50 mM Tris-HCl buffer, pH 8.5, containing 10 ␮M PMSF, 11c), anti-CD56 (Leu 19), and anti-CD14 (Leu M3; Becton Dickinson, 100 ␮M bacitracin, 1 mM benzamidine, 1 ␮g/ml aprotinin, 1 ␮g/ml leu- Mountain View, CA); anti-CD16 (B73.1) and anti-MHC class I (W6/32; peptin, 1 ␮g/ml pepstatin, and 5 ␮g/ml soybean trypsin inhibitor. Cells provided by Dr. G. Trinchieri, Wistar Institute, Philadelphia, PA); anti- were lysed by sonication, and protein concentrations were determined us- CD56 (B159.5.2; provided by Dr. B. Perussia, Thomas Jefferson Univer- ing the Bio-Rad protein assay (Bio-Rad, Richmond, CA). Sixty to one ␤ sity, Jefferson Cancer Center, Philadelphia, PA); anti- 1 integrin subunit hundred micrograms of the whole cell lysate (or an appropriate amount of (4B4, purchased from Coulter Immunology (Hialeah, FL), and TS2/16, supernatant) was added to 250 ␮l of reaction buffer (50 mM Tris-HCl (pH by guest on September 25, 2021 provided by Dr. M. Hemler, Dana-Farber Cancer Institute (Boston, MA)); 8.5), 5 mM CaCl , 5 mM MgCl , and 0.1% fatty acid-free BSA) containing ␣ ␣ 2 2 anti- 4 (HP2/1) and anti- 5 (SAM-1) purchased from Immunotech (Mar- 1 ␮M radiolabeled PC vesicles (prepared by sonication in 50 mM Tris- Ј seille, France); and affinity-purified F(ab )2 of goat anti-mouse Ig (GAM) HCl, pH 8.5, in an ice bath for 5 min at 5 W and 80% output) and incubated purchased from Cappel Laboratories (Malvern, PA). at 37°C for 1 h. The reaction was stopped by the addition of 250 ␮lof The sPLA2 inhibitor p-bromophenacyl bromide ( pBPB) and the PLD chloroform/methanol/acetic acid (4/2/1). Two hundred and fifty microliters inhibitor 2,3-diphosphoglycerate (DPG) as well as purified bacterial PLD ␮ ␮ of H2O, 250 l of CHCl3, and 100 l of 4 M KCl were added, and the from Streptomyces chromofuscus (S. chromofuscus) were obtained from mixture was centrifuged at 4000 rpm for 5 min to separate the organic from Sigma (St. Louis, MO). the aqueous phase. The former was dried under nitrogen, resuspended in Two differently radiolabeled phosphatidylcholine (PC) were used to 200 ␮l of chloroform, and applied to a silica gel TLC plate (Merck, Darm- prepare PC vesicles for PLA2 assay; both were purchased from DuPont- stadt, Germany) in duplicate with an automatic applicator (Linomat IV, New England Nuclear (Boston, MA): L-␣-1-stearoyl-2-arachidonyl[arachi- 14 Camag, Muttenz, Switzerland). Samples were chromatographed in chloro- donyl-1- C]PC (sp. act., 56 mCi/mmol) and L-␣-dipalmitoyl-[choline- 14 form/methanol/acetic acid/water (100/60/16/8) to separate the major prod- methyl- C]PC (sp. act., 56 mCi/mmol). Phosphatidylinositol bisphosphate ucts of PLA activity, i.e., arachidonic acid (AA) and lyso-PC. Standards ␣ 14 2 (PtdIns(4,5)P2; L- -1-stearoyl-2-arachidonyl [arachidonyl-1- C]PtdIns; were chromatographed in parallel. The radioactive spots were visualized by sp. act., 20–50 mCi/mmol) and inositol-1,4,5-trisphosphate (Ins(1,4,5)P3) autoradiography, scraped from the plate, and counted in a beta counter. 3H radioreceptor assay were also obtained from DuPont-New England PLA2 activity was quantitated by the release of AA or lyso-PC from PC Nuclear (Boston, MA). and was expressed as the percent increase above the basal level of un- RIA kits for leukotriene (LT) C4/D4/E4, LTB4, and platelet-activating 3 treated samples. In some experiments cell lysates were pretreated with the factor (PAF) as well as [9,10- H]oleic acid were obtained from Amersham reducing agent DTT (Sigma) for 15 min at 4°C before performing the (Aylesbury, U.K.). assay.

PLC activity was assayed using radiolabeled PC or PtdIns(4,5)P2 ves- Preparation of human NK cell cultures icles as substrate in 50 mM Tris-HCl buffer, pH 7.0, containing 10 ␮M PMSF, 100 ␮M bacitracin, 1 mM benzamidine, 1 ␮g/ml aprotinin, 1 ␮g/ml NK cells were obtained by coculturing nylon nonadherent human PBMC leupeptin, 1 ␮g/ml pepstatin, and 5 ␮g/ml soybean trypsin inhibitor. Assay ϫ 5 from buffy coats (4 10 /ml) with irradiated (3000 rad) RPMI 8866 cells conditions were the same as those described for PLA activity measure- ϫ 5 2 (1 10 /ml) at 37°C in a humidified 5% CO2 atmosphere for 10 days as ment. To separate the parent phospholipid from PLC activity products, previously described (26). On day 10 the cell population was routinely ϩ ϩ Ϫ Ϫ such as DAG, the organic phase was chromatographed in petroleum ether/ 80–90% CD56 CD16 CD3 CD14 as assessed by immunofluorescence diethyl ether/acetic acid (70/30/1). PLC activity was quantitated by the and cytofluorometric analysis. In some experiments contaminating T cells release of DAG from PC or PtdIns(4,5)P2 and was expressed as the percent were eliminated by negative panning on plastic dishes. Anti-CD5-pre- increase above the basal level of untreated samples. Ins(1,4,5)P produc- ϫ 6 3 treated cells (10–20 10 ) were added to plastic petri dishes coated with tion was measured by competitive radiobinding assay kit. GAM (10 ␮g/ml) and incubated at 4°C for 2 h. The nonadherent cells were gently poured off. The resulting NK cell population was Ͼ90% ϩ ϩ Ϫ Ϫ CD16 CD56 CD3 CD14 as assessed by immunofluorescence and Serine protease secretion assay cytofluorometric analysis. NK cell populations containing no more than Ϫ Ϫ ϩ ␤ 2% contaminating CD16 CD56 CD3 cells were used for the present Purified anti-CD16 (B73.1), anti-CD56 (B159.5.2) or anti- 1 (4B4) mAbs experiments. and ECM (human plasma fibronectin (FN; Life Technologies, 2066 INTEGRIN-MEDIATED INHIBITION OF PLD AND DEGRANULATION IN NK CELLS

production (Fig. 1). PEt production in cells stimulated with GAM cross-linked anti-integrin mAbs alone was superimposable to that observed in cells stimulated with anti-CD56 control mAb alone (data not shown). Overall, these data suggest that the engagement of both FN re- ceptors expressed by NK cells down-modulates CD16-triggered PLD activity. ␤ 1 integrin cross-linking inhibits CD16-induced sPLA2, but not cPLA2 nor PI-PLC, activity in human NK cells Since CD16 engagement triggers the activation of other enzymatic activities involved in phospholipid turnover, such as PI-PLC and

both cytosolic and secretory PLA2 (19, 20, 22), we then investi- ␤ gated whether the 1 integrin inhibitory effect was selective for ␣ ␤ ␣ ␤ PLD or affected other phospholipases. Human NK cells were stim- FIGURE 1. 4 1 and 5 1 integrin cross-linking inhibits CD16-in- duced PLD activation in human NK cells. [3H]oleic acid-radiolabeled NK ulated with anti-CD16 mAb plus GAM in the presence or the ab- ␤ cells were stimulated with saturating doses of anti-CD16, anti-CD56, anti- sence of simultaneous cross-linking with anti- 1 or anti-CD56 ␤ ␣ ␣ 1, anti- 4, and anti- 5 mAbs, either alone or in combination, followed by control mAb; PI-PLC and cPLA2 enzymatic activities in the cell cross-linking with GAM for 5 min in the presence of 0.5% ethanol. Lipids lysates were then tested by measuring DAG or AA release from Downloaded from were then separated by TLC and the percentage of counts per minute in PEt radiolabeled PtdIns (4,5)P2 or arachidonyl-PC vesicles, respective- with respect to the total counts per minute in phospholipids was quantified ly; in addition, Ins (1,4,5)P3 production was measured by compet- by liquid scintillation counting as a measure of PLD activity. Results are itive radioreceptor binding assay. As shown in Fig. 2, neither DAG expressed as the percent increase in PEt production above the basal level release nor Ins(1,4,5)P3 production stimulated by CD16 engage- of untreated samples (0.4 pmol/mg protein). Results are representative of ␤ one of four independent experiments. ment was affected by simultaneous 1 cross-linking, suggesting

␤ http://www.jimmunol.org/ that CD16-induced PI-PLC activation is not modulated by 1 in- ␤ tegrins. Similarly, 1 integrin coengagement did not affect CD16- Grand Island, NY), and its 120- and 40-kDa proteolytic fragments (Chemi- stimulated DTT-insensitive AA release, indicating that CD16-me- con, Temecula, CA) were coimmobilized in flat-bottom tissue culture 24- ␤ diated cPLA2 activation is not influenced by 1 integrin ligation well plates (Costar, Cambridge, MA) by overnight incubation at 4°C. (Fig. 3A). PI-PLC activity and Ins(1,4,5)P3 generation as well as Plates were washed twice with cold PBS and blocked with PBS containing ␤ 1% BSA for 30 min at room temperature. Two hundred and fifty micro- cPLA2 activity in cells stimulated with GAM cross-linked anti- 1 liters of an NK cell suspension (8 ϫ 106 cells/ml) in RPMI 1640, 10 mM mAb alone were superimposable to those observed in cells stim- HEPES, and 1 mg/ml BSA were then plated for5hat37°C, and super- ulated with anti-CD56 control mAb alone (data not shown). natants were recovered by centrifugation at 100 ϫ g for 5 min. In some ␤ We then tested the effect of 1 integrin coengagement on CD16- experiments, cells were treated with DPG, ethanol, or wortmannin for 15 by guest on September 25, 2021 min at 37°C before being plated. Inhibitors were also present throughout induced sPLA2 activity. Human NK cells were incubated with anti- the test and did not affect cell viability, as tested by trypan blue exclusion. CD16 (B73.1) or anti-␤1 (4B4) mAb, either alone or in combina- In other experiments purified bacterial PLD from S. chromofuscus (25–100 tion, and then stimulated with GAM. Secretory PLA2 enzymatic U/ml) was added to the cell suspension immediately before seeding. activity was then tested in both the supernatant and the cell lysate Lymphocyte-specific serine protease activity in the supernatants was ␣ by measuring DTT-sensitive lyso-PC release from radiolabeled measured using the N- -benzyloxycarbonyl-L-lysine thiobenzyl ester ␤ (BLT; Calbiochem, La Jolla, CA) synthetic substrate as previously de- arachidonyl-PC vesicles. As shown in Fig. 3B, 1 integrin cross- scribed (22). Results were expressed as a percentage of the total cellular linking abolished both extracellular and cell-associated sPLA2 en- enzyme content after subtracting the spontaneous release, which did not zymatic activities induced by CD16 engagement, whereas simul- exceed 15%. taneous stimulation with the anti-CD56 control mAb had no effect. ␤ Similar results were obtained using a different anti- 1 (TS2/16) Results mAb (data not shown). CD16-induced sPLA inhibition was also ␤ 2 1 integrin cross-linking inhibits CD16-induced PLD activation ␣ ␣ observed using anti- 5 (SAM-1) or, to a lesser extent, anti- 4 in human NK cells (HP2/1) mAbs, suggesting a role for both FN receptors (Fig. 3B). ␤ Recent evidence has shown that 1 integrin ligation inhibits CD3- Secretory PLA2 activity in cells stimulated with GAM cross-linked induced T cell proliferation by interfering with membrane phos- anti-integrin mAbs alone was superimposable to that observed in pholipid metabolism, namely by decreasing PA and DAG produc- cells stimulated with anti-CD56 control mAb alone (data not tion (5). Since CD16 ligation stimulates a PLD activity that results shown). ␤ in PA production (21), we investigated whether 1 engagement We then measured by ELISA the amount of sPLA2 protein se- might modulate CD16-induced PLD activation in human NK cells. creted in the extracellular medium of stimulated cells and found 3 ␤ To this purpose, [ H]oleic acid-radiolabeled human NK cells were that simultaneous cross-linking with anti- 1 (4B4) mAb pro- ␤ incubated with saturating doses of anti-CD16 (B73.1) and anti- 1 foundly inhibited CD16-stimulated sPLA2 release in the superna- (4B4) mAb, alone or in combination, and then stimulated by GAM tant (Fig. 3C), suggesting that the inhibition of enzymatic activity cross-linking in the presence of ethanol. After stimulation, radio- might be due to the inhibition of protein secretion in the extracel- labeled PEt production was analyzed as a measure of PLD activity lular compartment. in vivo. As shown in Fig. 1, CD16-induced PEt production was We have recently demonstrated that both cPLA2 and sPLA2 ␤ completely inhibited by 1 integrin simultaneous cross-linking, contribute to CD16-induced PAF and AA metabolite generation in whereas anti-CD56 (B159.5.2) control mAb did not exert any ef- human NK cells in the early (1–5 min) and the late (5–40 min) fect. Similar results were observed using the TS2/16 mAb, recog- phases, respectively (22). Consistent with a selective inhibition of ␤ ␤ nizing a different, adhesion-stimulating, 1 epitope (data not the sPLA2 enzyme, 1 integrin cross-linking partially inhibited ␣ ␣ shown). Anti- 5 (SAM-1) and, to a lesser extent, anti- 4 (HP2/1) CD16-induced peptido-LT production at 5 min and prevented fur- mAb exerted a similar inhibiting effect on CD16-stimulated PEt ther accumulation at later time points (10 min) as assessed by RIA The Journal of Immunology 2067

␤ FIGURE 2. 1 integrin cross-linking does not affect CD16-stimulated PI-PLC enzymatic activity or Ins(1,4,5)P3 generation in human NK cells. Cells were stimulated with saturating doses of anti-CD16, anti-␤ , and anti-CD56 control mAb, either alone or in combination, followed by cross-linking with

1 Downloaded from GAM for 1 min, and PI-PLC enzymatic activity and Ins(1,4,5)P3 generation were then tested. A, PI-PLC activity in the cell lysate, evaluated as DAG release from radiolabeled PtdIns(4,5)P2 vesicles. B, Ins(1,4,5)P3 production, measured by competitive radiobinding assay. Results are expressed as the percent ␮ Ϯ increase in DAG release from PtdIns(4,5)P2 vesicles above the basal level of untreated samples (1.9 mol/mg protein/h; A) or net Ins(1,4,5)P3 production SE (picomoles per 106 cells) after subtracting the basal level of untreated samples (0.8 pmol/106 cells; B). Results are representative of one of three independent experiments.

(Fig. 3D). A similar degree of ␤ -mediated inhibition was also Taken together, these data indicate that ␤ integrin engagement

1 1 http://www.jimmunol.org/ observed for CD16-stimulated PAF and LTB4 production (data not in human NK cells selectively inhibits CD16-stimulated PLD ac- shown). tivation as well as sPLA2 protein secretion and enzymatic activity, by guest on September 25, 2021

␤ FIGURE 3. Effects of 1 integrin cross-linking on CD16-stimulated cPLA2 and sPLA2 enzymatic activity, sPLA2 protein release, and peptido-LT ␤ ␣ ␣ production. Human NK cells were stimulated with saturating doses of anti-CD16, anti-CD56, anti- 1, anti- 4, and anti- 5 mAbs, either alone or in combination, followed by cross-linking with GAM. Cell lysates were then tested for PLA2 enzymatic activity, and supernatants were assayed for sPLA2 protein release and enzymatic activity as well as for peptido-LT (LTC4,-D4, and -E4) production. A, cPLA2 activity in DTT-pretreated cell lysates after 1 min of stimulation, evaluated as AA release from radiolabeled arachidonyl-PC vesicles. B, sPLA2 activity in cell lysates (Pellet) and supernatants (SN) after 5 min of stimulation, evaluated as lyso-PC release from radiolabeled dipalmitoyl-PC vesicles. C, sPLA2 protein release in the supernatant, evaluated by ELISA. D, Peptido-LT accumulation in the supernatant, evaluated by RIA. Results are expressed as the percent increase in AA or lyso-PC release from

PC vesicles above the basal level of untreated samples (2 and 4 pmol/mg protein/h for AA and lyso-PC, respectively; A and B), micrograms of sPLA2 protein per 100 ␮g of protein in the cell lysate after subtracting the basal release of untreated samples (1.5 ␮g/100 ␮g protein; C), and picograms of peptido-LTs per milligram of protein in the cell lysate (D). Results are representative of one of three independent experiments. 2068 INTEGRIN-MEDIATED INHIBITION OF PLD AND DEGRANULATION IN NK CELLS

leaving unaffected CD16-dependent PI-PLC or cPLA2 activation. ␣ ␤ ␣ ␤ Both 4 1 and 5 1 FN receptors expressed on NK cells are in- volved in this process. Moreover, sPLA2 inhibition results in the partial inhibition of CD16-stimulated peptido-LT production.

CD16-stimulated sPLA2 secretion and enzymatic activity are dependent on PLD activation in human NK cells We then investigated whether there was a relationship between ␤ PLD and sPLA2 inhibition mediated by 1 integrins. To this pur- pose, we stimulated human NK cells with anti-CD16 (B73.1) mAb plus GAM in the presence or the absence of selective pharmaco- logical inhibitors of either sPLA2 ( pBPB) (22) or PLD (DPG) (33, 34). After stimulation, both enzymatic activities were tested as described above. Secretory PLA2 inhibition by pBPB had no de- tectable effect on CD16-triggered PLD activation (data not shown). Conversely, competitive inhibition of PLD activity by

DPG inhibited both sPLA2 protein release in the supernatant (Fig. 4A) and its enzymatic activity (Fig. 4B), in either cell-associated or Downloaded from extracellular form, in a dose-dependent manner, suggesting that

CD16-mediated PLD activation might be upstream of sPLA2 re- lease and activation. These results were further confirmed by the use of ethanol, which diverts PA generation to the production of the inactive metabolite PEt; as in the case of DPG, ethanol inhib- ited sPLA2 protein secretion as well as its extracellular or cell- http://www.jimmunol.org/ associated enzymatic activity (data not shown).

Together, these data indicate that CD16-induced sPLA2 release and activation in human NK cells are PLD dependent and suggest ␤ that 1 integrin-mediated inhibition of CD16-stimulated sPLA2 secretion and activity might be secondary to PLD inhibition.

FIGURE 4. PLD involvement in CD16-stimulated sPLA2 protein re- lease and enzymatic activity in human NK cells. Cells were pretreated with PLD involvement in CD16-induced NK cell granule exocytosis DPG for 15 min and stimulated with saturating doses of anti-CD16 or PLD enzymatic activity has been implicated in membrane traffick- anti-CD56 control mAb followed by cross-linking with GAM for 10 min, by guest on September 25, 2021 and both sPLA protein release and enzymatic activity were then tested. A, ing and granule secretion in several cell systems (35); in addition, 2 sPLA protein release in the supernatant, evaluated by ELISA. B, sPLA we have observed that PLD inhibition results in the inhibition of 2 2 activity in cell lysates (Pellet) and supernatants (SN), evaluated as lyso-PC CD16-stimulated sPLA2 protein secretion in human NK cells. release from radiolabeled dipalmitoyl-PC vesicles. Results are expressed as Thus, we investigated whether PLD enzymatic activity might be ␮ micrograms of sPLA2 protein per 100 g of protein in the cell lysate after involved in CD16-stimulated NK cell degranulation. To this pur- subtracting the basal release of untreated samples (1.2 ␮g/100 ␮g protein; pose, human NK cells were incubated with either DPG, a compet- A) and the percent increase in lyso-PC release from PC vesicles above the itive inhibitor of PLD, or wortmannin, a fungal metabolite that basal level of untreated samples (3.5 pmol/mg protein/h; B). Results are inhibits PLD activation (36, 37), and were stimulated with anti- representative of one of three independent experiments. CD16 mAb cross-linking. PLD enzymatic activity and BLT-ester- ase release in the supernatant of stimulated cells were then ana- ␤ lyzed. In experiments dealing with granule exocytosis, NK cells 1 integrin cross-linking inhibits CD16-stimulated, but not were stimulated by plastic-immobilized anti-CD16 mAbs, since PMA/ionomycin-stimulated, BLT-esterase release from human GAM cross-linked mAbs were unable to induce detectable degran- NK cells ulation (data not shown). As shown in Fig. 5, DPG inhibited both Recent evidence has shown that ECM components are able to PLD enzymatic activity and BLT-esterase release in a dose-depen- modulate receptor-stimulated granule exocytosis in several cell ␤ dent fashion; PLD enzymatic activity was completely inhibited at systems (3). The above reported evidence of a 1 integrin-medi- a DPG concentration of 5 mM, while BLT-esterase release was ated inhibition of a biochemical event (PLD activation) involved in inhibited by 25% at a DPG concentration of 5 mM and almost CD16-stimulated NK cell degranulation prompted us to investigate ␤ completely at 10 mM. Similarly, wortmannin completely inhibited whether 1 integrin ligation might modulate CD16-dependent PLD activation at a concentration of 30 nM, while complete inhi- granule exocytosis. To this purpose, human NK cells were stim- ␤ bition of BLT-esterase release required a higher inhibitor concen- ulated by plastic-coimmobilized anti-CD16 (B73.1) and anti- 1 tration (300 nM). These data suggest that PLD activation plays an (4B4) mAbs. After stimulation, BLT-esterase enzymatic activity in important, although not exclusive, role in CD16-triggered NK cell the supernatant was quantified as a measure of granule content ␤ degranulation. These results were further confirmed by the obser- secretion. As shown in Fig. 6A, 1 integrin coengagement partially vation that ethanol, which inhibits the PLD pathway by leading to inhibited CD16-induced degranulation, whereas anti-CD56 the production of the “false product” PEt instead of PA, also in- (B159.5.2) control mAb costimulation did not affect CD16-depen- ␣ ␤ ␣ ␤ hibited CD16-stimulated BLT-esterase release in a dose-dependent dent secretion. More importantly, 4 1 and 5 1 receptor engage- fashion (data not shown). ment by plastic-immobilized natural ligands, either human plasma Taken together, these data indicate that a PLD-dependent path- FN or its proteolytic fragments (40 and 120 kDa), resulted in a way is involved in CD16-mediated NK cell granule exocytosis. similar inhibition of CD16-induced NK cell degranulation (Fig. The Journal of Immunology 2069

FIGURE 5. Effect of PLD inhibition on CD16- stimulated PLD activity and BLT-esterase release in human NK cells. Human NK cells, radiolabeled (A and C) or not (B and D) with [3H]oleic acid, were pretreated with the indicated dose of either DPG or wortmannin for 15 and 30 min, respec- tively. A and C, PLD activity, evaluated as radio- active PEt production after 5 min of stimulation with GAM cross-linked anti-CD16 (solid circles) or anti-CD56 control mAb (open circles); results are expressed as indicated in Fig. 1. B and D, BLT- esterase activity released in the supernatant after 5 h of stimulation with plastic-immobilized anti- CD16 (solid circles) or anti-CD56 control mAb (open circles), assessed by colorimetric assay. Re- sults are expressed as the percentage Ϯ SE of total Downloaded from cellular BLT-esterase content released in the su- pernatant after subtracting the basal release of un- treated samples. Results are representative of one of three independent experiments. http://www.jimmunol.org/

6B). BLT esterase release from cells stimulated with plastic-im- molecular levels of interference between integrin- and other re- ␤ mobilized anti- 1 mAb or FN alone was superimposable to that ceptor-generated signals are not clearly established. ␤ obtained in cells stimulated with anti-CD56 control mAb or BSA Here we report the novel finding that cross-linking of 1 inte- ␣ ␤ ␣ ␤ alone (data not shown). grins, namely 4 1 and 5 1 FN receptors, specifically inhibits ␤ We then tested whether 1 integrin engagement selectively in- CD16-stimulated PLD activation in human NK cells. Moreover, ␤ terfered with CD16-induced degranulation by using pharmacolog- our data indicate that PLD inhibition by either 1 integrin engage- ical stimuli, such as PMA and ionomycin, which stimulate NK cell ment or pharmacological agents functionally results in the inhibi- by guest on September 25, 2021 secretion through pathways different from that elicited by CD16 tion of CD16-elicited NK cell granule exocytosis. To our knowl- (38). As shown in Fig. 6C, no difference in PMA and/or ionomy- edge this is the first report of a role for PLD in NK cell cin-stimulated BLT-esterase release was observed in the presence degranulation. Bonnema et al. have previously demonstrated that ␤ of plastic-immobilized anti- 1 or anti-CD56 control mAb. cytotoxic granule secretion triggered via CD16 is inhibited by ␤ To further strengthen the correlation between 1 integrin-in- wortmannin, a fungal metabolite that inhibits PI 3-kinase by bind- duced inhibition of both PLD and granule exocytosis, we investi- ing irreversibly to its p110 catalytic subunit (38). Wortmannin, gated whether exogenously added PLD might restore lytic granule however, has been used extensively in other cell systems to inhibit ␤ secretion in NK cells stimulated by CD16 and 1 simultaneous PLD activation induced by several stimuli (36, 37), and we have cross-linking. To this purpose, human NK cells were stimulated demonstrated that it also inhibits CD16-stimulated PLD activation ␤ with plastic-coimmobilized anti-CD16 (B73.1) and anti- 1 (4B4) in human NK cells (Fig. 5C). On these bases, we can hypothesize mAbs in the presence or the absence of escalating doses of purified that PLD might function as a downstream effector of PI 3-kinase bacterial PLD from S. chromofuscus. As shown in Fig. 6D, CD16- along the signaling pathway that links CD16 engagement to the ␤ stimulated BLT-esterase release was inhibited by 30% by 1 si- release of NK cytotoxic granules. Consistent with a specific inter- multaneous cross-linking and was returned to normal levels in the ference of integrin-generated signals on a CD16-triggered bio- ␤ ␤ presence of exogenously added purified PLD, suggesting that 1- chemical pathway that involves both PI 3-kinase and PLD, 1 mediated PLD inhibition may be responsible for the observed de- integrin cross-linking did not affect PMA- and/or ionomycin-in- crease in granule exocytosis. Exogenous PLD addition to unstimu- duced secretion, which has been demonstrated to be protein kinase lated cells did not cause BLT esterase release above the basal level C dependent and PI 3-kinase independent (38). Although the mo- (data not shown). lecular mechanisms by which PI 3-kinase may control PLD activ- ␤ Taken together, these data indicate that 1 integrin simultaneous ity are presently unknown, the recent observation that cross-linking selectively inhibits CD16-mediated granule exocy- PtdIns(3,4,5)P3 directly stimulates the activity of cloned, purified tosis in human NK cells without affecting PMA- and/or ionomy- PLDs provides a reasonable link between these two pathways (39). cin-induced degranulation. Moreover, reversal of this effect by ex- Several lines of evidence imply that PLD activation plays a ␤ ogenous PLD suggests that the 1 integrin-mediated decrease pivotal role in agonist-dependent secretion in different cell types. in granule exocytosis may be attributable to endogenous PLD PLD activation and exocytosis share common requirements, and inhibition. alterations in PLD regulation are linked to resistance of the secre- tory pathway to protein traffic-disrupting agents, such as brefeldin Discussion A (40). More direct approaches, using primary alcohols to divert Although integrin-mediated negative regulation of specific cell PA generation into PEt production and exogenous PLD to increase functions has been reported in several cell systems (4–16), the PA levels, have also provided evidence that PLD and its product 2070 INTEGRIN-MEDIATED INHIBITION OF PLD AND DEGRANULATION IN NK CELLS Downloaded from

␤ FIGURE 6. 1 integrin-mediated inhibition of CD16-stimulated, but not of PMA/ionomycin-induced, NK cell degranulation and reversal by exogenous

␤ http://www.jimmunol.org/ PLD addition. Human NK cells were stimulated with plastic-immobilized anti-CD16, anti- 1 or anti-CD56 control mAb, and ECM proteins (FN and its proteolytic fragments, 40 and 120 kDa), either alone or in combination (A, B, and D) or with PMA and ionomycin (Iono) at the indicated doses (C). After 5 h of incubation at 37°C, supernatants were tested for BLT esterase activity in a colorimetric assay. A, Inhibition of CD16-induced BLT esterase release ␤ by escalating doses of coimmobilized anti- 1 (solid circles) or anti-CD56 (open circles) control mAb. B, Inhibition of CD16-induced BLT esterase release ␤ by coimmobilized FN and its proteolytic fragments (40 and 120 kDa). C, Effect of anti- 1 (solid bars) or anti-CD56 (white hatched bars) control mAb on ␤ BLT esterase release from cells stimulated by either coimmobilized anti-CD16 or escalating doses of PMA and ionomycin. D, Anti- 1-mediated inhibition of BLT esterase release from cells stimulated by coimmobilized anti-CD16 in the presence of escalating doses of exogenously added purified PLD from S. chromofuscus; the indicated doses of PLD were added simultaneously to cell seeding on mAb-coated plates. Results are expressed as the percent Ϯ SE of inhibition of CD16-stimulated BLT esterase release (A, B, and D) and the percentage Ϯ SE of total cellular BLT esterase content released in the supernatant after subtracting the basal release of untreated samples (B). Results are representative of one of five independent experiments. by guest on September 25, 2021

PA are involved in transport vesicle assembly and secretion in tosis was only observed after stimulation with plastic-immobilized several cell types (35, 41–43). anti-CD16 mAbs, PLD activation occurred with either GAM ␤ ␤ CD16-induced sPLA2 activity was also inhibited upon 1 inte- cross-linked or plastic-immobilized mAbs. However, 1 integrin grin simultaneous cross-linking, functionally resulting in a de- coengagement inhibited PLD activation in both conditions (M. crease in peptido-LT, LTB4, and PAF production. Since pharma- Milella, unpublished observations). In addition, inhibition of de- cological inhibition of PLD activity also resulted in sPLA2 granulation required the coimmobilization of both anti-CD16 and ␤ ␤ inhibition, it is conceivable that in CD16-stimulated NK cells 1 anti- 1 mAbs on the same plastic surface, since pretreatment with ␤ integrin-mediated inhibition of sPLA2 may be a consequence of soluble or GAM cross-linked anti- 1 mAb did not modify the abil- PLD inhibition. Although the molecular mechanisms by which ity of plastic-immobilized CD16 to induce granule exocytosis

PLD may control sPLA2 activation in human NK cells are pres- (data not shown). ently unknown, several lines of evidence indicate that a secretory Interestingly, NK cell adhesion to FN or its 120- and 40-kDa event may play a pivotal role in this process. PLD inhibition by fragments exerted a similar inhibitory effect on degranulation (Fig. ␤ either 1 integrin engagement or pharmacological agents results in 6B), further adding to the physiological relevance of these find- the parallel inhibition of both sPLA2 protein secretion and enzy- ings. Similar results have been obtained in human eosinophils, matic activity; since enzyme release in the extracellular medium, whose adhesion to laminin- and FN-coated wells inhibits degran- where it finds millimolar concentrations of Ca2ϩ, may be per se ulation in a concentration-dependent and secretagogue-specific sufficient for its activation (44), the inhibition of sPLA2 release manner (13). Conversely, FN has been shown to costimulate TCR- may explain the observed inhibition of its enzymatic activity. In induced degranulation in mouse CTL clones (46–48). One possi- addition, although sPLA2 storage in NK granules has not been ble explanation for these discrepant findings is that in such an ␣ ␤ demonstrated, enzyme secretion parallels granule exocytosis (22, experimental system adhesion to FN was mediated by the v 3 45), and enzymatic activity is inhibited by several agents that in- vitronectin receptor, as suggested by the blocking effect of the terfere with NK cell degranulation with different mechanisms, RMV-7 mAb (46). ␤ such as the PI 3-kinase inhibitor wortmannin, the mitogen-acti- In previous studies we demonstrated that 1 integrin cross-link- vated protein kinase kinase inhibitor PD 098059, or PLD inhibitors ing on human NK cells enhances both natural and Ab-dependent (22) (M. Milella et al., unpublished observations; Fig. 4). cytotoxicity (28). Although apparently discrepant, these findings Together, our data provide evidence that simultaneous cross- might be explained by the existence of diverse cytotoxic mecha- ␤ linking of 1 integrins results in a decrease in CD16-triggered NK nisms (49, 50) that can be differently regulated by integrins and cell granule exocytosis due to the selective inhibition of a PLD- suggest that integrins may interact with CD16 in several ways, dependent pathway. It is worth noting that while granule exocy- contributing to the fine regulation of NK cell functions. An even The Journal of Immunology 2071 more intriguing explanation is that the integrin-mediated attenua- References tion of NK cell degranulation might be instrumental in the poten- 1. Defilippi, P., A. Gismondi, A. Santoni, and G. Tarone. 1997. Integrin structure tiation of their cytotoxicity. Indeed, after delivering the lethal hit to and function. In Signal Transduction by Integrins. Springer-Verlag, Heidelberg, a susceptible target, an NK cell may recycle, become inactivated p. 1. 2. Clark, E. A., and J. S. Brugge. 1995. Integrins and signal transduction pathways: (due to depletion of critical effector molecules or postreceptor de- the road taken. Science 268:233. sensitization), or undergo activation-induced cell death (18). Both 3. Defilippi, P., A. Gismondi, A. Santoni, and G. Tarone. 1997. Integrin regulation inactivation and apoptosis decrease the ability of the affected NK of hematopoietic cell function. In Signal Transduction by Integrins. Springer- Verlag, Heidelberg, p. 165. cell to mediate subsequent killing of additional targets. Decreased 4. Groux, H., S. Huet, H. Valentin, D. Pham, and A. Bernard. 1989. Suppressor degranulation might limit both effector molecule depletion and the effects and cyclic AMP accumulation by the CD29 molecule of CD4ϩ lympho- release of granzymes, which are involved in the apoptotic process, cytes. Nature 339:152. 5. Ticchioni, M., C. Aussel, J.-P. Breittmayer, S. Manie´, C. Pelassy, and A. Bernard. thereby preserving NK cell recycling ability and potentiating their 1993. Suppressive effect of T cell proliferation via the CD29 molecule. The CD29 ␤ cytotoxicity. Preliminary data indicate that 1 integrin cross-link- mAb 1 “K20” decreases diacylglycerol and phosphatidic acid levels in activated ing protects CD16-stimulated NK cells from apoptosis (S. Mor- T cells. J. Immunol. 151:119. 6. Kent Teague, T., and B. W. McIntyre. 1994. MAb 18D3 triggering of integrin ␤1 rone, unpublished observations). Similarly, eosinophil adhesion to will prevent but not terminate proliferation of human T cells. Cell Adhes. Com- FN and laminin prolongs their in vitro survival while attenuating mun. 2:169. 7. Kitani, A., N. Nakashima, T. Matsuda, B. Xu, S. Yu, T. Nakamura, and degranulation (13). T. Matsuyama. 1996. T cells bound by vascular cell adhesion molecule-1/CD106 ␤ The molecular mechanisms by which 1 integrin cross-linking in synovial fluid in rheumatoid arthritis: inhibitory role of soluble vascular cell adhesion molecule-1 in T cell activation. J. Immunol. 156:2300.

might interfere with CD16-induced PLD activation remain to be Downloaded from 8. Dransfield, I., C. Caban˜as, J. Barrett, and N. Hogg. 1992. Interaction of leukocyte established. Although CD16-induced PLD activation in human integrins with ligand is necessary but not sufficient for function. J. Cell Biol. ϩ NK cells is strongly dependent upon extracellular Ca2 and is 116:1527. accompanied by polyphosphoinositide breakdown (21), ␤ inter- 9. Hurley, R. W., J. B. McCarty, and C. M. Verfaille. 1995. Direct adhesion to bone 1 marrow stroma via fibronectin receptors inhibits hematopoietic progenitor pro- ference at these levels seems unlikely. Indeed, we have previously liferation. J. Clin. Invest. 96:511. ␤ 10. Pellat-Deceunynck, C., M. Amiot, N. Robillard, J. Wijdenes, and R. Bataille. demonstrated that 1 integrins synergize with CD16 in the induc- 2ϩ 1996. CD11a-CD18 and CD102 interactions mediate human myeloma cell tion of a Ca influx from the extracellular compartment (27), and growth arrest induced by CD40 stimulation. Cancer Res. 56:1909. http://www.jimmunol.org/ we provide here evidence that CD16-stimulated PI-PLC enzymatic 11. Lundell, B. I., J. B. McCarty, N. L. Kovach, and C. M. Verfaille. 1996. Activa- ␣ ␤ activity and Ins(1,4,5)P generation are not affected by ␤ simul- tion-dependent 5 1 integrin-mediated adhesion to fibronectin decreases prolif- 3 1 eration of chronic myelogenous leukemia progenitors and K562 cells. Blood 87: taneous cross-linking (Fig. 2). Several other biochemical events 2450. involved in the regulation of mammalian PLD have been identified 12. Bathia, R., J. B. McCarty, and C. M. Verfaille. 1996. Interferon-␣ restores normal ␤ in different cell types, including activation of protein kinase C 1 integrin-mediated inhibition of hematopoietic progenitor proliferation by the marrow microenvironment in chronic myelogenous leukemia. Blood 87:3883. isozymes, ADP-ribosylation factors, and small G proteins, such as 13. Kita, H., S. Horie, and G. J. Gleich. 1996. Extracellular matrix proteins attenuate RhoA, Rac1, and Cdc42Hs (42, 43). In addition, both activation and degranulation of stimulated eosinophils. J. Immunol. 156:1174. PtdIns(4,5)P and PtdIns(3,4,5)P strongly stimulate PLD by di- 14. Marth, T., and B. L. Kelsall. 1997. Regulation of interleukin-12 by complement 2 3 receptor 3 signaling. J. Exp. Med. 185:1987. by guest on September 25, 2021 rectly activating the purified enzyme (39). Most of these pathways 15. Hwang, S.-M., C. A. Lopez, D. E. Heck, C. R. Gardner, D. L. Laskin, share a common molecular theme, which is the need for translo- J. D. Laskin, and D. T. Denhardt. 1994. Osteopontin inhibits induction of nitric oxide synthase gene expression by inflammatory mediators in mouse kidney ep- cation to specific membrane sites. Thus, a possible explanation for ithelial cells. J. Biol. Chem. 269:711. our findings is that integrin-induced cytoskeletal reorganization 16. Jackson, S. P., S. M. Schoenwaelder, Y. Yuan, I. Rabinowitz, H. H. Salem, and may interfere with the effective recruitment of specific signaling C. A. Mitchell. 1994. Adhesion receptor activation of phosphatidylinositol 3-ki- nase: von Willebrand factor stimulates the cytoskeletal association and activation components involved in CD16-triggered PLD activation. Consis- of phosphatidylinositol 3-kinase and pp60c-src in human platelets. J. Biol. Chem. tent with a pivotal role of the cytoskeleton in integrin-mediated 269:27093. negative signaling, Bathia et al. have recently demonstrated that 17. Trinchieri, G. 1989. Biology of natural killer cells. Adv. Immunol. 47:187. 18. Leibson, P. J. 1997. Signal transduction during activation: in- actin polymerization inhibitors, such as cytochalasin D, com- side the mind of a killer. Immunity 6:665. ␤ pletely reverse the anti- 1 mAb-mediated inhibition of chronic 19. Azzoni, L., M. Kamoun, T. W. Salcedo, P. Kanakaraj, and B. Perussia. 1992. ␥ ␥ myelogenous leukemia progenitor proliferation (12). In this regard Stimulation of Fc RIIIA results in phospholipase C- 1 tyrosine phosphorylation and pp56lck activation. J. Exp. Med. 156:1745. ␤ we have recently demonstrated that 1 integrin engagement in hu- 20. Ting, A. T., L. M. Karnitz, R. A. Schoon, R. T. Abraham, and P. J. Leibson. 1992. man NK cells results in the phosphorylation of both PYK-2 (be- Fc␥ receptor activation induces the tyrosine phosphorylation of both phospho- ␥ ␥ longing to the focal adhesion kinase subfamily) and the PYK-2- lipase C (PLC)- 1 and PLC- 2 in natural killer cells. J. Exp. Med. 156:1751. 21. Balboa, M. A., J. Balsinde, J. Aramburu, F. Mollinedo, and M. Lopez-Botet. associated cytoskeletal protein paxillin (30). The formation of 1992. Phospholipase D activation in human natural killer cells through the Kp43 multicomponent signaling complexes in parallel to the structural and CD16 surface antigens takes place by different mechanisms: involvement of the phospholipase D pathway in tumor necrosis factor ␣ synthesis. J. Exp. Med. complexes at the site of integrin adhesion may therefore be rele- 176:9. ␤ vant to 1 integrin-mediated interference with CD16-generated 22. Milella, M., A. Gismondi, P. Roncaioli, L. Bisogno, G. Palmieri, L. Frati, signals. This hypothesis is currently under investigation. M. G. Cifone, and A. Santoni. 1997. CD16 cross-linking induces both secretory ␤ and extracellular signal-regulated kinase (ERK)-dependent cytosolic phospho- In summary, our data indicate that in human NK cells 1 inte- lipase A2 (PLA2) activity in human natural killer cells: involvement of ERK, but grins may interact with CD16-initiated pathways in different and not PLA2, in CD16-triggered granule exocytosis. J. Immunol. 158:3148. 23. Cifone, M. G., P. Roncaioli, L. Cironi, C. Festuccia, A. Meccia, S. D’Alo`, complex ways depending on the specific function. We also present D. Botti, and A. Santoni. 1997. NKR-P1A stimulation of arachidonate-generating the first evidence of a role for PLD activation in the secretory enzymes in rat NK cells is associated with granule release and cytotoxic activity. pathway elicited by CD16 engagement in NK cells. J. Immunol. 159:309. 24. Gismondi, A., S. Morrone, M. J. Humphries, M. Piccoli, L. Frati, and A. Santoni. 1991. Human natural killer cells express VLA-4 and VLA-5, which mediate their adhesion to fibronectin. J. Immunol. 146:384. 25. Gismondi, A., F. Mainiero, S. Morrone, G. Palmieri, M. Piccoli, L. Frati, and Acknowledgments A. Santoni. 1992. Triggering through CD16 or phorbol esters enhances adhesion We thank Dina Milana, Anna Maria Bressan, Patrizia Birarelli, Alessandro of NK cells to laminin via very late antigen 6. J. Exp. Med. 176:1251. 26. Mainiero, F., A. Gismondi, M. Milella, S. Morrone, G. Palmieri, M. Piccoli, Procaccini, Antonio Sabatucci, and Gasperina De Nuntiis for their expert L. Frati, and A. Santoni. 1994. Long-term activation of NK cells results in mod- technical assistance. ulation of ␤1-integrin expression and function. J. Immunol. 152:446. 2072 INTEGRIN-MEDIATED INHIBITION OF PLD AND DEGRANULATION IN NK CELLS

27. Palmieri, G., A. Gismondi, R. Galandrini, M. Milella, A. Serra, R. De Maria, and terization of two alternately spliced forms of phospholipase D1: activation of the A. Santoni. 1997. Interaction of natural killer cells with extracellular matrix in- purified enzymes by phosphatidylinositol 4,5-bisphosphate, ADP-ribosylation duces early intracellular signalling events and enhances cytotoxic functions. Nat. factor, and Rho family monomeric GTP-binding proteins and protein kinase C- Immun. 15:147. alpha. J. Biol. Chem. 272:6 3860. 28. Palmieri, G., A. Serra, R. De Maria, A. Gismondi, M. Milella, M. Piccoli, 40. Ktistakis, N. T., H. A. Brown, M. G. Waters, P. C. Sternweis, and M. G. Roth. ␣ ␤ ␣ ␤ L. Frati, and A. Santoni. 1995. Cross-linking of 4 1 and 5 1 fibronectin re- 1996. Evidence that phospholipase D mediates ADP ribosylation factor-depen- ceptors enhances natural killer cell cytotoxic activity. J. Immunol. 155:5314. dent formation of Golgi coated vesicles. J. Cell Biol. 134:295. 29. Gismondi, A., M. Milella, G. Palmieri, M. Piccoli, L. Frati, and A. Santoni. 1995. 41. Cockroft, S. 1992. G-protein-regulated phospholipases C, D and A -mediated Stimulation of protein tyrosine phosphorylation by interaction of NK cells with 2 ␣ ␤ ␣ ␤ signalling in . Biochim. Biophys. Acta 1113:135. fibronectin via 4 1 and 5 1. J. Immunol. 154:3128. 30. Gismondi, A., L. Bisogno, F. Mainiero, G. Palmieri, M. Piccoli, L. Frati, and 42. Singer, W. D., H. A. Brown, and P. C. Sternweis. 1997. Regulation of eukaryotic ␤ phosphatidylinositol-specific phospholipase C and phospholipase D. Annu. Rev. A. Santoni. 1997. Proline-rich tyrosine kinase-2 activation by 1 integrin fi- bronectin receptor cross-linking and association with paxillin in human natural Biochem. 66:475. killer cells. J. Immunol. 159:4729. 43. Exton, J. H. 1997. New developments in phospholipase D. J. Biol. Chem. 272: 31. Mainiero, F., A. Gismondi, A. Soriani, M. Cippitelli, G. Palmieri, J. Jacobelli, 15579. M. Piccoli, L. Frati, and A. Santoni 1998. Integrin-mediated Ras-extracellular 44. Kudo, I., M. Murakami, S. Hara, and K. Inoue. 1993. Mammalian non-pancreatic ␥ ␥ regulated kinase (ERK) signalling regulates interferon- (IFN- ) production in phospholipases A . Biochim. Biophys. Acta 117:217. human natural killer cells. J. Exp. Med. 188:1267. 2 32. Bligh, E. G., and W. J. Dyer. 1959. A rapid method of total lipid extraction and 45. Ting, A. T., R. A. Schoon, R. T. Abraham, and P. J. Leibson. 1992. Interaction purification. Can. J. Biochem. Physiol. 37:911. between protein kinase C-dependent and G protein-dependent pathways in the 33. Kanaho, Y., Y. Nakai, M. Katoh, and Y. Nozawa. 1993. The phosphatase inhib- regulation of natural killer cell granule exocytosis. J. Biol. Chem. 267:23957. itor 2,3-diphosphoglycerate interferes with phospholipase D activation in rabbit 46. Takahashi, K., T. Nakamura, H. Adachi, H. Yagita, and K. Okumura. 1991. neutrophils. J. Biol. Chem. 268:12492. Antigen-independent T cell activation mediated by a very late activation antigen- 34. Kusner, D. J., C. F. Hall, and L. S. Schlesinger. 1996. Activation of phospho- like extracellular matrix receptor. Eur. J. Immunol. 21:1559. lipase D is tightly coupled to the of Mycobacterium tuberculosis or 47. Ybarrondo, B., A. M. O’Rourke, A. A. Brian, and M. F. Mescher. 1994. Con- Downloaded from opsonized zymosan by human . J. Exp. Med. 184:585. tribution of lymphocyte function-associated-1/intercellular adhesion molecule-1 35. Roth, M. G., and P. C. Sternweis. 1997. The role of lipid signalling in constitutive binding to the adhesion/signaling cascade of cytotoxic T lymphocyte activation. membrane traffic. Curr. Opin. Cell Biol. 9:519. J. Exp. Med. 179:359. 36. Reinhold, S. L., S. M. Prescott, G. A. Zimmerman, and T. M. McIntyre. 1990. 48. Ostergaard, H. L., and M. A. Eileen. F. 1995. Fibronectin induces phosphoryla- Activation of human phospholipase D by three separable mechanisms. tion of a 120-kDa protein and synergizes with the T cell receptor to activate FASEB J. 4:208. cytotoxic T cell clones. Eur. J. Immunol. 25:252. 37. Ward, S. G., C. H. June, and D. Olive. 1996. PI 3-kinase: a pivotal pathway in T cell activation? Immunol. Today 17:187. 49. Henkart, P. A., M. S. Williams, and H. Nakajima. 1995. Degranulating cytotoxic

38. Bonnema, J. D., L. M. Karnitz, R. A. Schoon, R. T. Abraham, and P. J. Leibson. T inflict multiple damage pathways on target cells. Curr. Top. Mi- http://www.jimmunol.org/ 1994. stimulation of phosphatidylinositol-3 kinase in natural killer crobiol. Immunol. 198:75. cells is associated with protein kinase C-independent granule release and cell- 50. Eischen, C. M., J. D. Schilling, D. H. Lynch, P. H. Krammer, and P. J. Leibson. mediated cytotoxicity. J. Exp. Med. 180:1427. 1995. Fc receptor-induced expression of on activated NK cells facil- 39. Hammond, S. M., J. M. Jenco, S. Nakashima, K. Cadwallader, Q. Gu, S. Cook, itates cell-mediated cytotoxicity and subsequent autocrine NK cell apoptosis. Y. Nozawa, G. D. Prestwich, M. A. Frohman, and A. J. Morris. 1997. Charac- J. Immunol. 156:2693. by guest on September 25, 2021