Fgr and Hck Neutrophils Requires the Src Family Kinases Adhesion-Dependent Degranulation Of

Adhesion-Dependent Degranulation of Neutrophils Requires the Src Family Kinases Fgr and Hck

This information is current as Attila Mócsai, Erzsébet Ligeti, Clifford A. Lowell and of September 25, 2021. Giorgio Berton J Immunol 1999; 162:1120-1126; ; http://www.jimmunol.org/content/162/2/1120 Downloaded from

References This article cites 36 articles, 23 of which you can access for free at: http://www.jimmunol.org/content/162/2/1120.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• 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

*average by guest on September 25, 2021

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. Adhesion-Dependent Degranulation of Neutrophils Requires the Src Family Kinases Fgr and Hck1

Attila Mo´csai,2*† Erzse´bet Ligeti,† Clifford A. Lowell,‡ and Giorgio Berton3*

Polymorphonuclear neutrophils (PMN) adherent to integrin ligands respond to inflammatory mediators by reorganizing their cytoskeleton and releasing reactive oxygen intermediates. As Src family tyrosine kinases are implicated in these responses, we investigated their possible role in regulating degranulation. Human PMN incubated on fibrinogen released lactoferrin in response to TNF-␣ and this response was inhibited by PP1, a Src family tyrosine kinase inhibitor. This drug had no effect on lactoferrin secretion induced by PMA, an adhesion-independent agonist of PMN degranulation. However, PP1 blocked secretion in PMN plated on plain tissue culture plastic, a surface inducing PMN spreading in the absence of any stimulus. Double knockout hck؊/؊fgr؊/؊ PMN adherent to collagen or fibrinogen failed to release lactoferrin in response to TNF-␣ but responded to PMA ؊/؊ ؊/؊ as wild-type PMN. Degranulation induced by spreading over tissue culture plastic was also defective in hck fgr PMN. Downloaded from Defective adhesion-dependent degranulation required the absence of both kinases, because single knockout fgr؊/؊ or hck؊/؊ PMN responded as wild-type cells. Analysis of lactoferrin secretion in hck؊/؊fgr؊/؊ or PP1-treated, suspended PMN showed that Src kinases are not implicated in degranulation dependent on activation of protein kinase C or increase in intracellular free Ca2؉ but may play a role in the response to FMLP of cytochalasin B-treated PMN. These findings identify a role for Src family kinases in a signaling pathway leading to granule-plasma membrane fusion and suggest that Fgr and Hck would be targets for pharmacological control of adhesion-dependent degranulation in the inflammatory site. The Journal of Immunology, 1999, 162: 1120–1126. http://www.jimmunol.org/

olymorphonuclear neutrophils (PMN)4 adherent to extra- hesion and spreading of PMN are tightly coupled to an increase in cellular matrix proteins or endothelial counterreceptors re- tyrosine phosphorylation of several proteins, and tyrosine kinase P spond to inflammatory mediators, such as cytokines and inhibitors block adhesion-dependent generation of ROI by PMN chemoattractants, releasing reactive oxygen intermediates (ROI) (7, 8). The Src family tyrosine kinases Fgr, Hck, and Lyn, as well and granule constituents (1–5). These secretory responses are as p72syk, have been recently implicated in signaling from adhe- tightly coupled with a rearrangement of the cytoskeleton that dic- sion receptors (16–19). tates spreading of the cell over the adhesive surface (5–8). Dif- Signals involved in triggering degranulation by adherent PMN ferent adhesion receptors have been implicated in induction of are poorly understood. The evidence that protein tyrosine phos- by guest on September 25, 2021 ␤ ␤ PMN responses. For example, both 2 and 3 integrins have been phorylation (7, 16) and the Src family kinases Fgr and Hck (18) demonstrated to regulate PMN responses (2, 3, 6, 9–11). The re- play a critical role in ROI generation prompted us to investigate lease of ROI and degranulation in the context of PMN-endothelial whether adhesion-dependent degranulation is also regulated by the cell interactions plays a central role in the development of inflam- same signaling pathways. In this report, we show that adhesion- matory reactions and certainly contributes to tissue damage asso- dependent release of lactoferrin is blocked by a newly described ciated with inflammation (12, 13). tyrosine kinase inhibitor the effect of which has been reported to be In the last few years, a great progress has been made in the selective for Src family tyrosine kinases (20). Investigations with understanding of signaling from adhesion receptors (14, 15). Ad- PMN isolated from mice with the double deficiency of Fgr and Hck allowed us to demonstrate that these kinases play an essential role in signaling for adhesion-dependent degranulation. These re- † *Institute of General Pathology, University of Verona, Verona, Italy; Department of sults provide the first direct evidence that Src family kinases func- Physiology and Laboratory of Cellular and Molecular Physiology, Semmelweis Uni- versity of Medicine, Budapest, Hungary; and ‡Department of Laboratory Medicine, tion in a signaling pathway leading to granule-plasma membrane University of California, San Francisco, CA 94143 fusion. Received for publication March 26, 1998. Accepted for publication October 6, 1998. The costs of publication of this article were defrayed in part by the payment of page Materials and Methods charges. This article must therefore be hereby marked advertisement in accordance PMN isolation with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by grants from the Italian Association for Cancer Research Human PMN were isolated from buffy coats of healthy volunteers as pre- and Minister of University and Scientific Research of Italy/University of Verona viously described (9, 16). Wild-type (C57BL/6), p59/61hck,orp58c-fgr- (Research Program: Inflammation: Biology and Clinic) to G.B.; National Institutes of deficient single knockout mice and double knockout mice deficient in both Health Grants DK 50267 and HL54476 to C.A.L.; and grants from the Hungarian p59/61hck and p58c-fgr were described previously (21). Bone marrow of National Research Fund (OTKA) and the Swiss National Science Foundation 6–12-wk-old mice of both sexes were used in this study. Murine PMN (7UNPJ048717) to E.L. were isolated essentially as described (18) except that a simplified three- 2 Recipient of a short term fellowship from the European Molecular Biology layer gradient was utilized to separate neutrophils from other bone marrow Organization. cells (i.e., bone marrow cells suspended in Ca2ϩ-/Mg2ϩ-free HBSS sup- 3 Address correspondence and reprint requests to Dr. Giorgio Berton, Istituto di Pa- plemented with 0.1% BSA were layered on the top of a 62/81% two-layer tologia Generale, Universita`di Verona, Strada Le Grazie, 37134 Verona, Italy. E-mail Percoll (Pharmacia, Uppsala, Sweden) gradient, and after centrifugation, address: [email protected] PMN were harvested from the 62/81% interface). At the end of the prep- 4 Abbreviations used in this paper: PMN, polymorphonuclear neutrophils; ROI, aration, PMN were suspended in ice-cold HBSS containing 0.5 mM CaCl2, reactive oxygen intermediates; TNF, TNF-␣; PKC, protein kinase C; CB, 1 mM MgCl2,and5mMD-glucose (HBSS) and kept in ice until use. For ϩ cytochalasin B. experiments in which the Mg2 dependence of the PMN response was

investigated, cells were isolated as above described but resuspended in 2ϩ 2ϩ Results Ca /Mg -free HBSS supplemented with 5 mM EDTA and, after 10 min TNF-stimulated lactoferrin release by adherent human and of incubation, washed and suspended in HBSS either with or without 1 mM Mg2ϩ (17). murine PMN is blocked by the tyrosine kinase inhibitor PP1 Human adherent PMN can be stimulated to mobilize lactoferrin- Coating of tissue culture plastic plates with proteins containing speciﬁc granules in response to granulocyte-macro- Flat bottom polystyrene tissue culture plates with 96 wells (Greiner, Frick- phage-CSF, FMLP, or the Ca2ϩ ionophore A23187 (5). As shown enhausen, Germany) were left untreated or were covered with rat collagen, in Fig. 1A, TNF is also an effective agonist of lactoferrin secretion human ﬁbrinogen, or FCS as described (18). After coating, the plates were washed once with PBS and once with HBSS. in adherent human PMN. As previously reported with other stimuli (5), TNF-induced degranulation by adherent PMN is delayed in its Cell stimulation onset and prolonged up to 60 min of incubation (Fig. 1A). In ac- Human and murine PMN were washed at 4°C and resuspended in ice-cold cord with previous studies (4, 23–25), we could not detect release HBSS at 1 ϫ 106 or 2 ϫ 106 cells/ml, respectively. For adhesion assays, of the primary granule marker ␤-glucuronidase in response to TNF cell suspensions (100 ␮l/well), added or not with 10 ␮M PP1 (Calbiochem- (data not shown). Additionally, we also found that the response to Novabiochem Int., La Jolla, CA) were dispensed in protein-coated plates TNF depended on adhesion (25, 26), since suspended PMN failed and prewarmed for 10 min at 37°C before addition of 20 ng/ml human or murine TNF-␣ (TNF, Peprotech, London, U.K.), 100 ng/ml PMA (Sigma, to release lactoferrin following TNF stimulation while such cells St. Louis, MO), or diluent alone. Since the TNF stock solution was stored did respond to PMA or ionomycin (data not shown; see Fig. 6 for in the presence of BSA, all the other samples were supplemented with the data with murine PMN). same concentration of BSA. After incubation at 37°C for the indicated

As a ﬁrst approach to elucidate signals involved in degranula- Downloaded from times, samples were transferred to ice-cold polypropylene tubes and cen- tion by adherent PMN, we examined whether the compound PP1, trifuged at 8000 ϫ g for 1 min at 4°C. Supernatants were used for the determination of granule marker release into the extracellular space. For which was recently described as a relatively selective inhibitor of suspension assays, 100-␮l aliquots of a 2 ϫ 106/ml cell suspension were Src family tyrosine kinases (20), affected TNF-stimulated lacto- dispensed in polypropylene tubes and either preincubated or not with 10 ferrin secretion. As shown in Fig. 1A,10␮M PP1 completely ␮ M cytochalasin B (CB) and/or PP1 for 10 min at 37°C. PMN were stim- inhibited lactoferrin secretion in response to TNF. Importantly, ulated with 1 ␮M FMLP (Sigma), 1 ␮M ionomycin, or 100 ng/ml PMA, PP1 inhibited only the adhesion-dependent response; in fact,

and after 10 min samples were transferred to ice-cold polypropylene tubes http://www.jimmunol.org/ and processed as described above for adhesion assays. PMA-induced degranulation was unaffected by the drug (Fig. 1B). Previous evidence established a strict link between adhesion- Determination of lactoferrin release dependent stimulation of PMN functions and cell spreading (5–8). Release of the speciﬁc granule marker lactoferrin was measured by an In line with this evidence, we observed that PP1 completely ELISA using anti-human lactoferrin Abs (Sigma). Supernatants of human blocked PMN spreading on ﬁbrinogen in response to TNF, but not 2Ϫ or murine PMN were diluted 4- or 20-fold, respectively, in 50 mM CO3 / PMA (Fig. 2). In addition, generation of superoxide anion, which ␮ HCO3 buffer (pH 9.6). A total of 100 l of the diluted supernatants or of known concentrations of human lactoferrin were added to Nunc Maxisorp is also stimulated as PMN are induced to spread by TNF, was F96 (Nunc, Roskilde, Denmark) immunoplate wells and incubated over- blocked by PP1 (L. Fumagalli and G. Berton, unpublished obser-

night at 4°C. All subsequent steps were conducted at room temperature and vation). It is unlikely that PP1 acts by inhibiting TNF receptor by guest on September 25, 2021 separated by several washings. Nonspecific binding sites were blocked signaling because it also blocked lactoferrin secretion that oc- with PBS supplemented with 0.5% BSA and 0.5% Tween 20 (blocking curred when PMN were plated on plain tissue culture plastic, a solution). Plates were then treated with affinity-purified rabbit anti-human lactoferrin (Sigma; dilution, 1/500) followed by peroxidase-conjugated anti- surface that can rapidly induce PMN spreading, in the absence of rabbit Ab (Amersham, Little Chalfont, U.K.; dilution, 1/5000), both dis- any stimulus (Fig. 1C). solved in blocking solution. Color was developed by the o-phenylenedi- The described inhibitory effects of PP1 on adhesion-dependent amine/H2O2 system. Absorbance of the wells were read at 490 nm with a degranulation were also observed using PMN isolated from mouse Packard Argus 300 microplate reader, and the lactoferrin concentration was calculated using the human lactoferrin calibration curve. With both human bone marrow (Fig. 1D). As seen with human cells, murine PMN and murine samples, no signal was detected if the anti-lactoferrin Ab was released lactoferrin when incubated on collagen in the presence of substituted with an irrelevant rabbit antiserum. The reactivity of the anti- TNF or on plain tissue culture plastic in the absence of any stim- human lactoferrin Ab with murine lactoferrin was determined by Western ulus (see also Fig. 3). PP1 inhibited murine PMN lactoferrin se- 8 blotting. Murine PMN suspended at 10 /ml in HBSS were incubated for 15 cretion in both conditions (Fig. 1D). The high degranulation re- min at 37°C in the presence or the absence of 100 ng/ml PMA. Cells were then spun down and the supernatant solubilized in 4ϫ concentrated SDS- sponse induced by PMA in mouse PMN was unaffected by PP1 PAGE sample buffer and boiled for 10 min. Samples were run on a 10% (see legend to Fig. 1D). As found with human PMN (Fig. 2), PP1, SDS-PAGE, electroblotted onto nitrocellulose sheets (Hybond C, Amer- in parallel with its inhibition of degranulation also completely sham), blocked, and probed with anti-human lactoferrin primary and per- blocked murine PMN spreading in response to TNF, but not PMA oxidase-labeled secondary Abs exactly as described for ELISA measurements. Signal was developed using an Amersham enhanced (data not shown). chemiluminescence reagent. The anti-human lactoferrin Ab recognized one single band at Ϸ78 kDa in supernatants of murine PMN (not shown). The Adhesion-dependent lactoferrin release is defective in double m.w. of this band and the fact that its intensity increased in the supernatants knockout hckϪ/ϪfgrϪ/Ϫ PMN of PMA-stimulated PMN confirmed this protein to be lactoferrin. Previous studies demonstrated that adhesion-dependent generation Measurement of ␤-glucuronidase release of superoxide anion is defective in hckϪ/ϪfgrϪ/Ϫ PMN (18). Hav- Release of the primary granule marker ␤-glucuronidase from the same ing established that adherence and spreading induce lactoferrin supernatants used for lactoferrin measurements was conducted essentially release in murine PMN and that the Src family selective inhibitor as described (22). PP1 blocks this response in both human and murine PMN (Fig. 1), Expression of the data we addressed whether deficiency of Fgr and Hck results in a defective adhesion-dependent degranulation. Wild-type PMN re- All the experiments were performed in duplicates or triplicates. Since data leased lactoferrin in response to TNF if incubated on collagen or, obtained with mouse samples could not be exactly quantified using standard curves done with human lactoferrin, lactoferrin release from murine albeit to a lower extent, on fibrinogen, while they responded poorly PMN is expressed as fold increase compared with the amount detected in if incubated on FCS (Fig. 3). In addition, wild-type PMN released supernatants of unstimulated PMN (see figure legends). lactoferrin when incubated on plain tissue culture plastic even in 1122 NEUTROPHIL ADHESION-DEPENDENT DEGRANULATION NEEDS Fgr AND Hck Downloaded from http://www.jimmunol.org/

FIGURE 1. The tyrosine kinase inhibitor PP1 blocks adhesion-dependent lactoferrin secretion by human and murine PMN. PMN were incubated on different surfaces for the times indicated as described in Materials and Methods. Stimuli (20 ng/ml TNF or 100 ng/ml PMA) were added after 10 min of preincubation with 10 ␮M PP1. Mean results Ϯ SD of triplicates assays of one representative experiment, which was reproduced three times, are reported. Data obtained with murine PMN (D) are expressed as fold increase relative to the amount of lactoferrin found in the supernatant of unstimulated PMN incubated for 30 min on collagen (see Materials and Methods). In the same set of experiments reported in D, PMA induced an increase of lactoferrin release by guest on September 25, 2021 by murine PMN of 6.16- Ϯ 0.66-fold in the absence, and 6.35- Ϯ 0.84-fold in the presence of PP1, respectively. the absence of TNF. Comparable results were previously obtained culture plastic was comparable in wild-type and single knockout by assaying superoxide anion release by wild-type PMN (18). Lac- PMN, but totally defective in the double knockout cells. toferrin release in response to TNF by hckϪ/ϪfgrϪ/Ϫ double knockout PMN was defective on all the surfaces tested. Impor- Adhesion-independent lactoferrin release occurs normally in tantly, the spontaneous lactoferrin release by PMN incubated on Ϫ/Ϫ Ϫ/Ϫ tissue culture plastic was also totally defective in hckϪ/ϪfgrϪ/Ϫ double knockout hck fgr PMN PMN. The findings described in Figs. 3–5 implicate Fgr and Hck in sig- As lactoferrin release in response to TNF was optimal when naling for adhesion-dependent degranulation. However, the defec- PMN were adherent to collagen we investigated degranulation by tive response of hckϪ/ϪfgrϪ/Ϫ PMN could be also due to the fact collagen-adherent PMN in more detail. As shown in Fig. 4A, lac- that Hck and Fgr play a more general role in signaling for degran- toferrin release by wild-type PMN adherent to collagen and stimulation and/or granule-plasma membrane fusion. As a first ap- ulated with TNF was delayed and maximal at about 30 min. We proach to demonstrate that Hck and Fgr are required for signaling did not detect any release of the primary granule marker ␤-glucu- from integrins, we addressed whether TNF-induced lactoferrin re- ronidase under these conditions. hckϪ/ϪfgrϪ/Ϫ double knockout lease by murine PMN is indeed dependent on adhesion. To this PMN did not release lactoferrin in response to TNF at any of the purpose, we exploited previous findings demonstrating that PMN time points tested. However, PMA induced release of lactoferrin to adhesion in response to TNF is absolutely dependent on the pres- a comparable extent in wild-type and double knockout PMN (Fig. ence of Mg2ϩ in the incubation medium (17). As shown in Fig. 6, 4B). Light microscopy observation confirmed previous findings in experiments in which lactoferrin release by murine PMN was (18) that hckϪ/ϪfgrϪ/Ϫ PMN did not spread in response to TNF investigated in suspension assays, TNF was unable to induce a but responded to PMA as well as wild-type cells. response whether Mg2ϩ was included or not in the assay medium. Because adhesion-dependent production of superoxide anion is However, wild-type, but not hckϪ/ϪfgrϪ/Ϫ PMN, released lacto- normal in single knockout fgrϪ/Ϫ or hckϪ/Ϫ PMN (18), we sought ferrin in response to TNF if incubated on collagen, and this re- to investigate whether the knockout of the two kinases is required sponse was strictly dependent on the presence of Mg2ϩ. Since to cause a total deficiency of adhesion-dependent lactoferrin re- Mg2ϩ is required for recognition of extracellular matrix proteins lease. As shown in Fig. 5, double knockout, but not single knock- by integrins, the data demonstrate that impairment of integrin-me- out fgrϪ/Ϫ or hckϪ/Ϫ, PMN were totally defective in their capa- diated adhesion in wild-type cells produced the same defect in bility to release lactoferrin in response to TNF when plated on degranulation seen in mutant PMN under normal culture condi- collagen. In addition, degranulation induced by spreading on tissue tions. This supports the model that the defect we found in The Journal of Immunology 1123

FIGURE 2. The tyrosine kinase inhibitor PP1 blocks TNF-induced, human PMN spreading over ﬁbrinogen. Human PMN were incubated in ﬁbrino- gen-coated wells for 30 min after 10 min of preincubation without (A, B, D) or with 10 ␮M PP1 (C, E). Cells were either unstimulated (A) or treated with 20 ng/ml TNF (B, C) or 100 ng/ml PMA (D, E). PP1 blocked cell spreading induced by TNF, but not that induced by PMA. Downloaded from http://www.jimmunol.org/

hckϪ/ϪfgrϪ/Ϫ PMN concerns an adhesion-dependent cell re- (PKC) and Ca2ϩ. However, we found that hckϪ/ϪfgrϪ/Ϫ PMN sponse. were indeed defective in the response to FMLP (Fig. 7B). This by guest on September 25, 2021 To test directly whether hckϪ/ϪfgrϪ/Ϫ PMN were able to release ﬁnding suggests a role of Src family kinases in signaling by FMLP lactoferrin in assays not requiring adhesion, we analyzed degran- and is concordant with the evidence that FMLP activates Src fam- ulation occurring in suspended PMN in response to different stim- ily kinases in PMN (Refs. 27–29, see Discussion). It must be, uli. As shown in Fig. 7A), lactoferrin release in standard suspen- however, noted that FMLP-induced lactoferrin secretion required sion assays in response to PMA or the Ca2ϩ ionophore ionomycin disruption of the actin-based cytoskeleton with CB, while PMA- is comparable in wild-type and hckϪ/ϪfgrϪ/Ϫ PMN. Hence, Hck and ionomycin-induced degranulation occurred also in the absence and Fgr do not regulate granule-plasma membrane fusion and are of this drug (Fig. 7). At present, we do not know whether Hck and dispensable for lactoferrin secretion dependent on protein kinase C Fgr are implicated in signaling by FMLP and/or in facilitation of FMLP signaling by cytoskeletal rearrangements induced by CB. To reinforce the evidence that Hck and Fgr are not implicated in PKC- and Ca2ϩ-dependent lactoferrin secretion, but play a role in FMLP ϩ CB-induced degranulation, we addressed whether the Src kinase inhibitor PP1 affected degranulation in suspension assays. As shown in Fig. 8A, PP1 inhibited the FMLP ϩ CB-induced degranulation, albeit at a higher concentration than that required to block adhesion-dependent degranulation (see Fig. 1). However, lactoferrin release induced by PMA or ionomycin was unaffected by PP1 (Fig. 8B). We conclude that Src family kinase inhibition by PP1 affects PMN degranulation in the same fashion as the knockout of hck and fgr, i.e., inhibits the response to FMLP ϩ CB but does not modify the PKC- and Ca2ϩ-induced response.

Discussion FIGURE 3. Adhesion-dependent lactoferrin (Lfr) secretion is defective Signaling from adhesion molecules in PMN involves Src family in hckϪ/ϪfgrϪ/Ϫ PMN. Wild-type (WT) and double knockout (H/F) PMN tyrosine kinases. In this report, we demonstrate that adhesion-de- were incubated on the indicated surfaces for 30 min in the presence or absence of 20 ng/ml TNF. Data are expressed as fold increase relative to pendent release of lactoferrin-containing speciﬁc granules is de- the amount of lactoferrin found in the supernatant of WT PMN incubated pendent on the activity of tyrosine kinases and on the expression of for 30 min on collagen (see Materials and Methods). Mean results Ϯ SD at least two members of the Src family, i.e., the Fgr and Hck of triplicates assays of one representative experiment, which was repro- proteins. PP1, which has been described as a selective inhibitor of duced three times, are reported. Src family kinases (20), inhibits adhesion-dependent lactoferrin 1124 NEUTROPHIL ADHESION-DEPENDENT DEGRANULATION NEEDS Fgr AND Hck

FIGURE 6. TNF-induced lactoferrin (Lfr) secretion is strictly dependent on adhesion and is defective in hckϪ/ϪfgrϪ/Ϫ PMN. Murine PMN were incubated with EDTA and resuspended in HBSS supplemented or not with1mMMg2ϩ as described in Materials and Methods. Cells were dis-

pensed in polypropylene tubes (Susp) or plated in collagen-coated wells Downloaded from (Adh) and incubated for 30 min in either the absence or the presence of 20 ng/ml TNF. Data are expressed as fold increase relative to the amount of lactoferrin found in the supernatant of WT PMN incubated for 30 min in the presence of Mg2ϩ. Mean results Ϯ SD of triplicate assays of one representative experiment, which was reproduced twice, are reported. http://www.jimmunol.org/ hckϪ/ϪfgrϪ/Ϫ PMN responded normally to PMA and a Ca2ϩ ionophore. These findings strongly argue that Fgr and Hck are critically involved in transducing signals through adhesion molecules, but degranulation can still be induced by activating PKC with phorbol esters or increasing intracellular free Ca2ϩ concentrations. Hence, FIGURE 4. Adhesion-dependent lactoferrin (Lfr) secretion is defective although Fgr and Hck were found to localize at least in part in in hckϪ/ϪfgrϪ/Ϫ PMN. Wild-type (WT) and double knockout (H/F) PMN PMN-specific and primary granules, respectively (30, 31), they are were incubated on collagen for the indicated times in the presence of 20 dispensable for PKC- and Ca2ϩ-induced degranulation. ng/ml TNF (A) or 100 ng/ml PMA (B). Mean results Ϯ SD of five inde- As shown for ROI generation (18), adhesion-dependent degran- by guest on September 25, 2021 pendent experiments are reported. Data are expressed as reported in Figs. ulation requires cell stimulation with a cytokine, such as TNF 1 and 3. (Refs. 4 and 23–26, and this article), or other agonists (4, 5). It is unlikely that Fgr and Hck are primarily involved in agonist signaling. In fact, stimulus-independent degranulation induced by release from both human and murine PMN. In the absence of Fgr spreading of PMN on tissue culture plastic was blocked by PP1 and Hck, PMN failed to mobilize specific granules after adhesion and was also defective in hckϪ/ϪfgrϪ/Ϫ PMN. Additionally, TNF- to different surfaces. PP1 does not affect specific granule exocy- Ϫ/Ϫ Ϫ/Ϫ ϩ induced up-regulation of CD11b/c is normal in hck fgr tosis in response to either PMA or a Ca2 ionophore. In addition, PMN (18). The evidence that PP1 blocks human and murine PMN spreading and that hckϪ/ϪfgrϪ/Ϫ fail to spread (18) suggests a strict link between PMN spreading and degranulation. Thus, it is most likely that signaling from adhesion receptors requires Src family tyrosine kinases to initiate cytoskeletal rearrangements leading to cell spreading and subsequent degranulation. Analyzing degranulation in suspension assays, we found that hckϪ/ϪfgrϪ/Ϫ PMN and PMN treated with the Src kinase inhibitor PP1 are defective in the response induced by FMLP in the presence of CB. These findings suggest that signaling by the FMLP receptor also involves Src family kinases. Indeed, several reports have implicated Src family kinases in signaling by FMLP in PMN (27–29), as well as by heterotrimeric GTP-binding protein-coupled receptors in other cell types (reviewed in Ref. 32). The data reported in Fig. 7 do suggest that Hck and Fgr play a critical role in FMLP signaling in PMN. Indeed, we have also found that activation of the respiratory burst by FMLP in CB-treated hckϪ/ϪfgrϪ/Ϫ FIGURE 5. Adhesion-dependent lactoferrin (Lfr) secretion is defective Ϫ Ϫ Ϫ Ϫ PMN is defective and that FMLP activates the Src family kinases in double, but not single knockout PMN. Wild-type (WT), hck / , fgr / , or double knockout (H/F) PMN were incubated on collagen for 30 min Hck and Fgr in CB-treated human PMN (L. Fumagalli, C. A. Low- either in the absence of any stimulus or in the presence of 20 ng/ml TNF. ell, and G. Berton, manuscript in preparation). Activation of PMN Mean results Ϯ SD of triplicate assays of one representative experiment, responses by FMLP in murine PMN is absolutely dependent on the which was reproduced twice, are reported. Data are expressed as reported disassembly of the actin-based cytoskeleton by cytochalasins. In a in Figs. 1 and 3. previous study (18) and the present one, we invariably found that The Journal of Immunology 1125 Downloaded from http://www.jimmunol.org/

FIGURE 7. Lactoferrin (Lfr) release by suspended hckϪ/ϪfgrϪ/Ϫ PMN FIGURE 8. PP1 inhibits lactoferrin (Lfr) release by suspended murine is not defective in response to PMA or ionomycin, but defective in re- PMN in response to FMLP, but not PMA or ionomycin (iono). PMN were sponse to FMLP. Wild-type (WT) or hckϪ/Ϫ fgrϪ/Ϫ double knockout (H/F) incubated and assayed as described in Fig. 7. In the experiments reported PMN were incubated in polypropylene tubes for 10 min in the absence or by guest on September 25, 2021 in B 20 ␮M PP1 was used. Data are expressed as reported in Fig. 7. Mean presence of 10 ␮M CB and then stimulated with 1 ␮M ionomycin and 100 results Ϯ SD of triplicate assays of one representative experiment, which ng/ml PMA or 1 ␮M FMLP. After 10 min of incubation, lactoferrin re- was reproduced twice, are reported. leased in the supernatant was assayed as described in Materials and Meth- ods. Mean results of triplicate assays of four independent experiments are reported. Data are expressed as fold increase relative to the amount of lactoferrin found in the supernatant of unstimulated WT PMN incubated in were unable to spread on different surfaces (18). As with degran- the absence of CB. ulation, ROI generation was also shown to be strictly correlated with spreading (18). Thus, it appears that loss of Src family kinase activity, either through the use of PP1 or in the knockouts, pri- hckϪ/ϪfgrϪ/Ϫ PMN display alterations in the organization of the marily affects PMN spreading which is in turn required for ROI actin cytoskeleton, which may contribute to their inability to generation and degranulation. spread over integrin ligands or even plain tissue culture plastic. To some extent, it is surprising that degranulation proceeds in Hence we cannot exclude that impairment of the response to coincidence with the extensive process of actin polymerization that FMLP ϩ CB results from cytoskeletal alterations present in double accompanies cell spreading. In fact, the block of actin polymer- mutant PMN. ization by cytochalasins has been used since the early 1970s to A clear hierarchy of mobilization of different granule compart- enhance PMN degranulation (see Ref. 34 for a review of early ments, i.e., secretory vesicles Ͼ gelatinase granules Ͼ specific studies). Additionally, at least with murine PMN, we could not granule (lactoferrin-containing) Ͼ azurophil granules, has been detect any significant release of lactoferrin in FMLP-stimulated demonstrated in human PMN (33). To our knowledge, it is not cells assayed in suspension if cells were not pretreated with CB known whether specific granules of murine PMN can be divided (Fig. 7). Cytoskeleton rearrangement is a highly dynamic process into gelatinase-containing and lactoferrin-containing subsets. (35). Thus, localized rearrangements of the actin cytoskeleton may Since hckϪ/ϪfgrϪ/Ϫ PMN are not defective in the capability to provide a driving force for membrane fusion between adjacent up-regulate CD11b/CD11c (18), which are also stored in secretory granules and plasma membrane. In the absence of Hck and Fgr vesicles (33), it is tempting to speculate that Fgr- and Hck-regu- activity, the signals to initiate the appropriate cytoskeletal lated adhesion selectively controls mobilization of distinct granule changes do not occur. Interestingly, members of the Rho sub- subsets. family of GTP-binding proteins that play a central role in actin Although PMN spreading can occur independently of degranu- polymerization have been described to regulate degranulation in lation (5), previous studies (5) and those described in this report mast cells (36–38). suggest that degranulation by adherent PMN requires cell spread- PMN degranulation is thought to play an important role in tissue ing. PP1 blocked spreading of both human and murine PMN in damage and development of inflammation (12, 13, 33). Impor- response to TNF or FMLP, but not PMA, and hckϪ/ϪfgrϪ/Ϫ PMN tantly, we recently found that hckϪ/ϪfgrϪ/Ϫ mice are markedly 1126 NEUTROPHIL ADHESION-DEPENDENT DEGRANULATION NEEDS Fgr AND Hck resistant to the lethal effect of high dose lipopolysaccharide injec- 18. Lowell, C. A., L. Fumagalli, and G. Berton. 1996. Deficiency of src-family ki- hck c-fgr tion, and this correlates with reduction in liver and renal damage nases p59/61 and p58 results in defective adhesion-dependent neutrophil functions. J. Cell Biol. 133:895. (39). Reduced production of reactive oxygen intermediate (18) and 19. Yan, S.R., and G. Berton. 1997. Signaling by adhesion in human neutrophils. release of granule constituents (this article) by PMN adherent to Activation of the p72syk tyrosine kinase and formation of protein complexes syk the vascular wall of inflamed tissues may explain resistance of containing p72 and Src family kinases in neutrophils spreading over fibrinogen. J. Immunol. 158:1902. Ϫ/Ϫ Ϫ/Ϫ hck fgr mice to endotoxic shock. Implication of the Src 20. Hanke, J. H., J. P. Gardner, R. L. Dow, P. S. Changelian, W. H. Brissette, family tyrosine kinases Fgr and Hck as regulators of adhesion- E. J. Weringer, B. A. Pollok, and P. A. Connelly. 1996. Discovery of a novel, dependent PMN degranulation suggests that they would be new potent, and Src-family-selective tyrosine kinase inhibitor: study of Lck- and Fyn- dependent T cell activation. J. Biol. Chem. 271:695. targets for pharmacological control of the inflammatory process. 21. Lowell, C. A., P. Soriano, and H. E. Varmus. 1994. Functional overlap in the src gene family: inactivation of hck and fgr impairs natural immunity. Gene Dev. 8:387. References 22. Moćsai, A., B. Banfi, A. Kapus, G. Farkas, M. Geiszt, L. Buday, A. Farago, and 1. Nathan, C. F. 1987. PMN activation on biological surfaces: massive secretion of E. Ligeti. 1997. Differential effects of tyrosine kinase inhibitors and an inhibitor hydrogen peroxide in response to products of macrophages and lymphocytes. of the mitogen activated protein kinase cascade on degranulation and superoxide J. Clin. Invest. 80:1550. production of human neutrophil granulocytes. Biochem. Pharmacol. 54:781. 2. Nathan, C., S. Srimal, C. Farber, E. Sanchez, L. Kabbash, A. Asch, J. Gailit, and 23. Klebanoff S. J., M. A. Vadas, J. M. Harlan, L. H. Sparks, J. R. Gamble, S. D. Wright. 1989. Cytokine-induced respiratory burst of human neutrophils: J. M. Agosti, and A. M. Waltersdorph. 1986. Stimulation of neutrophils by tumor dependence on extracellular matrix proteins and CD11/CD18 integrins. J. Cell necrosis factor. J. Immunol. 136:4220. Biol. 109:1341. 24. Ferrante A., M. Nandoskar, A. Waltz, D. H. B. Goh, and I. C. Kowanko. 1988. 3. Shappel, S. B., C. Toman, D. C. Anderson, A. A. Taylor, M. L. Entman, and Effects of tumor necrosis factor ␣ and interleukin-1 and ␣ and ␤ on human C. W. Smith. 1990. Mac-1 (CD11b/CD18) mediates adherence-dependent hy- neutrophil migration, respiratory burst and degranulation. Int. Arch. Allergy Appl.

drogen peroxide production by human and canine neutrophils. J. Immunol. 144: Immunol. 86:82. Downloaded from 2702. 25. Luedke, E. S., and J. L. Humes. 1989. Effect of tumor necrosis factor on granule 4. Richter, J., T. Andersson, and I. Olsson. 1989. Effect of tumor necrosis factor and release and LTB4 production in adherent human neutrophils. Agents Action 27: granulocyte/macrophage colony-stimulating factor on neutrophil degranulation. 451. J. Immunol. 142:3199. 26. Tennenberg, S. D., and J. S. Solomkin. 1990. Activation of neutrophils by ca- 5. Suchard, J., and L. A. Boxer. 1994. Exocytosis of a subpopulation of speciﬁc chectin/tumor necrosis factor: priming of N-formyl-methionyl-leucyl-phenylala- granules coincides with H2O2 production in adherent human neutrophils. J. Im- nine-induced oxidative responsiveness via receptor mobilization without degran- munol. 152:290. ulation. J. Leukoc. Biol. 47:217. 6. Nathan, C. F., and E. Sanchez. 1990. Tumor necrosis factor and CD11/CD18 (␤ ) 2 27. Ptasznik A., A. Traynor-Kaplan, and G. M. Bokoch. 1995. G protein-coupled http://www.jimmunol.org/ integrins act synergistically to lower cAMP in human PMNs. J. Cell Biol. 111: chemoattractant receptors regulate Lyn tyrosine kinase: Shc adapter protein sig- 2171. naling complexes. J. Biol. Chem. 270:19969. 7. Fuortes, M., W. W. Jin, and C. Nathan. 1993. Adhesion-dependent protein ty- 28. Gaudry, M., C. Gilbert, F. Barabe´, P. E. Poubelle, and P. H. Naccache. 1995. rosine phosphorylation in neutrophils treated with tumor necrosis factor. J. Cell Activation of Lyn is a common element of the stimulation of human neutrophils Biol. 120:777. by soluble and particulate agonists. Blood 86:3567. 8. Laudanna, C., F. Rossi, and G. Berton. 1993. Effect of inhibitors of distinct Ϫ 29. Welch, H., and I. Maridonneau-Parini. 1997. Lyn and Fgr are activated in distinct signalling pathways on PMN O2 generation in response to tumor necrosis factor-␣, and antibodies against CD18 and CD11a: evidence for a common and membrane fractions of human granulocytic cells. Oncogene 15:2021. unique pattern of sensitivity to wortmannin and protein tyrosine kinase inhibitors. 30. Gutkind, J. S., and K. C. Robbins. 1989. Translocation of the FGR protein- Biochem. Biophys. Res. Commun. 190:935. tyrosine kinase as a consequence of neutrophil activation. Proc. Natl. Acad. Sci. 9. Berton, G., C. Laudanna, C. Sorio, and F. Rossi. 1992. Generation of signals USA 86:8783.

activating PMN functions by leukocyte integrins: LFA-1 and gp150/95, but not 31. Mo¨hn, H., V. Le Cabec, S. Fischer, and I. Maridonneau-Parini. 1995. The src- by guest on September 25, 2021 CR3, are able to stimulate the respiratory burst of human PMNs. J. Cell Biol. family protein-tyrosine kinase p59hck is located on the secretory granules in hu- 116:1007. man neutrophils and translocates towards the phagosome during cell activation. 10. Zhou, M., and E. J. Brown. 1993. Leukocyte response integrin and integrin- Biochem. J. 309:657. associated protein act as a signal transduction unit in generation of a phagocyte 32. Gutkind, J. S. 1998. The pathways connecting G-protein coupled receptors to the respiratory burst. J. Exp. Med. 178: 1165. nucleus through divergent mitogen-activated protein kinase cascades. J. Biol. ϩ 11. Lowson, M. A., and F. R. Maxfield. 1995. Ca2 - and calcineurin-dependent Chem. 273:1839. recycling of an integrin to the front of migrating neutrophils. Nature 377:75. 33. Borregaard, N., and J. C. Cowland. 1997. Granules of the human neutrophilic 12. Albelda, S. M., C. W. Smith, and P. A. Ward. 1994. Adhesion molecules and polymorphonuclear leukocyte. Blood 89:3503. inflammatory injury. FASEB J. 8:504. 34. Klebanoff, S. J., and R. A. Clark. 1978. The Neutrophil: Function and Clinical 13. Berton, G., S. R. Yan, F. Fumagalli, and C. A. Lowell. 1996. Neutrophil activa- Disorders. North Holland Publishing Co., Amsterdam, The Netherlands. tion by adhesion: mechanisms and pathophysiological implications. Int. J. Clin. 35. Stossel, T. P. 1993. On the crawling of animal cells. Science 260:1086. Lab. Res. 26:160. 14. Clark, E. A., and J. S. Brugge. 1995. Integrins and signal transduction pathways: 36. Price, L. S., J. C. Norman, A. J. Ridley, and A. Koffer. 1995. The small GTPases the road taken. Science 268:233. Rac and Rho as regulators of secretion in mast cells. Curr. Biol. 5:68. 15. Yamada, K. M., and S. Miyamoto. 1995. Integrin transmembrane signaling and 37. Mariot, P., A. J. O’Sullivan, A. M. Brown, and P. E. R. Tatham. 1996. Rho cytoskeletal control. Curr. Opin. Cell Biol. 7:681. guanine nucleotide dissociation inhibitor protein (RhoGDI) inhibits exocytosis in ␤ mast cells. EMBO J. 15:6476. 16. Berton, G., F. Fumagalli, C. Laudanna, and C. Sorio. 1994. 2 integrin-dependent protein tyrosine phosphorylation and activation of the FGR protein tyrosine ki- 38. Prepens, U., I. Just, C. von Eichel-Streiber, and K. Aktories. 1996. Inhibition of nase in human neutrophils. J. Cell Biol. 126:1111. Fc⑀RI mediated activation of rat basophilic leukemia cells by Clostridium difficile 17. Yan, S. R., L. Fumagalli, and G. Berton. 1995. Activation of p58c-fgr and p53/ toxin B (monoglucosyltransferase). J. Biol. Chem. 271:7324. 56lyn in adherent human neutrophils: evidence for a role of divalent cations in 39. Lowell, C. A., and G. Berton. 1998. Resistance to endotoxic shock and reduced regulating neutrophil adhesion and protein tyrosine kinase activities. J. Inflam- neutrophil migration in mice deficient for the Src-family kinases Hck and Fgr. mation 45:297. Proc. Natl Acad. Sci. USA 95:7580.