JBC Papers in Press. Published on June 27, 2008 as Manuscript M710286200 The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.M710286200 THE CLEAVAGE OF NEUTROPHIL LEUKOSIALIN (CD43) BY CATHEPSIN G RELEASES ITS EXTRACELLULAR DOMAIN AND TRIGGERS ITS INTRAMEMBRANE PROTEOLYSIS BY PRESENILIN/γ-SECRETASE* Agnès Mambole1,2, Dominique Baruch2,3, Patrick Nusbaum1,2, Sylvain Bigot1,2, Misa Suzuki4, Philippe Lesavre1,2, Minoru Fukuda4 and Lise Halbwachs-Mecarelli1,2 From INSERM U8451, Université René Descartes 2, INSERM U7653, Paris, France, Cancer Research Center, Burnham Institute for Medical Research, La Jolla CA , USA4 Running head: CD43 processing by cathepsin G and γ-secretase Address correspondence to: Lise Halbwachs-Mecarelli, INSERM U 845, Hôpital Necker, 161 rue de Sèvres, 75015 Paris, France; Fax +33 144 49 02 90; e-mail: [email protected]; The highly negatively charged membrane The regulated proteolysis of transmembrane sialoglycoprotein leukosialin, CD43, is shed proteins represents an important mechanism of cell during neutrophil activation. This is generally functions modulation. The inflammation resolution thought to enhance cell adhesion. We here involves, for example, the shedding of cytokine describe two novel consequences of this receptors and adhesion molecules, which down- shedding, during neutrophil activation by regulates leukocyte adhesion to endothelium. These phorbol esters or by chemoattractants after regulations result from a decreased membrane TNF-α priming. CD43 proteolysis was expression of receptors and/or from the release of D o investigated by western blotting, using a soluble fragments, competing with their membrane w n polyclonal antibody to CD43 intracellular counterparts. The majority of shed proteins loa d domain. Our data emphasize the importance of a identified to date are cleaved by metalloproteinases ed juxtamembranous cleavage of about 50% of or by neutrophil-derived serine proteases (1). from membrane CD43 moleculesof CD43 by cathepsin Leukosialin, CD43 is the predominant cell surface http G. Indeed, its is inhibited by α1- sialoprotein of leukocytes (2) and has both anti- ://w antichymotrypsin and and cathepsin G inhibitor adhesive and adhesive properties. Its function has ww I and is reproduced by exogenous purified been mainly studied on lymphocytes, where CD43 .jb c cathepsin G. The resulting membrane-anchored behaves both as a negative-regulator of T cell .org C-Terminal Fragment, CD43-CTF, becomes proliferation and adhesion and as a positive by/ susceptible to presenilin/γ-secretase, which regulator of memory T cells trafficking (3,4). gu e releases CD43 intracytoplasmic domain: Although its expression is normally restricted to st o n preincubation with 3 different γ-secretase leukocytes, CD43 is present on colon carcinomas D e inhibitors, before PMN treatment by agonists or and on several nonhematopoietic cell lines (5 ,6). In cem these cell lines, CD43 is processed by a b by purified cathepsin G, results in the e accumulation of CD43-CTF. Since CD43 binds presenilin/γ-secretase-mediated Regulated Intra- r 11 E-selectin, we also investigated the effect of the membrane Proteolysis (RIP)1 (7). RIP refers to a , 201 8 soluble extracellular domain CD43s, released by sequential proteolysis of various type I membrane cathepsin G juxtamembranous cleavage, on neutrophil adhesion to endothelial cells. A 1 Abbreviations used are : α1PI, α1-proteinase recombinant CD43s-Fc fusion protein inhibited inhibitor (α1-antitrypsin) ; α1-CT , α1 anti- neutrophil E selectin-dependent adhesion to chymotrypsin ; BSA, bovine serum albumin ; endothelial cells under flow conditions, while it CD43-CTF, CD43 C-Terminal Fragment ; CD43s, had no effect on neutrophil static adhesion. soluble CD43 ; fMLP, N-formyl-L-methionyl-L- We thus propose that, in addition to its potential leucyl-L-phenylalanine ; Galgp, pro-adhesive role, CD43 proteolysis results in: i) galactoglycoprotein ; HBSS, Hanks balanced the release, by cathepsin G, of CD43 solution ; HUVEC, human umbilical vein extracellular domain, able to inhibit the endothelial cells ; MMP, matrix metalloproteinase ; adhesion of flowing neutrophils on endothelial NE, neutrophil elastase; PBS, phosphate buffered cells and thus to participate to the natural saline ; PMA, phorbol myristate acetate ; PMN, control of inflammation; ii) the release and/or polymorphonuclear neutrophil ; PMSF, phenyl- the clearance, by presenilin/γ-secretase, of CD43 methyl-sulfonyl-fluoride; PR3, proteinase 3 ; RIP, intracellular domain, thereby regulating CD43- Regulated Intra-membrane Proteolysis ; TNF, mediated signaling. Tumor necrosis factor 1 Copyright 2008 by The American Society for Biochemistry and Molecular Biology, Inc. proteins, including the amyloid precursor protein of also purchased from Calbiochem (Darmstadt, Alzheimer’s disease, the Notch receptor, CD44, E- Germany) and Athens Technology (Athens, cadherin…(8). RIP is initiated usually by a Georgia). Human IgGs were purified on protein A- metalloproteinase, which induces the shedding of sepharose from an AB blood group normal human the receptor ectodomain, followed by an serum. The CD43s-Fc fusion protein was produced intramembrane processing of the cell bound in COS cells, as described (22) and purified on fragment by a PS/γ secretase. This releases the protein A-sepharose. Tumor necrosis factor alpha, receptor intracellular domain in the cytoplasm. In TNF-α, was from PeproTech (Rocky Hill, NJ). the case of CD43 in cancer cell lines, this domain Recombinant human VCAM-1-Fc was from R&D translocates in the nucleus and causes the (Abingdon, UK). α1 anti-chymotrypsin (α1-CT), upregulation of various genes (9). Cathepsin G inhibitor I, and MG-132 were from CD43 has been recently described as a ligand for E- Calbiochem, Compound E from Alexis (Lausanne, selectin (10,11). Its expression on Switzerland), Tapi-0 and Tapi-1 from Peptides polymorphonuclear neutrophils (PMN) is ten fold International (Louisville, Kentucky), while higher than P-selectin glycoprotein ligand-1, PSGL- Marimastat, CH46644 and CH6631 were a kind gift 12 the main leukocyte ligand for P-selectin, which from Dr John Bird (Celltech Chiroscience, also binds E-selectin. In spite of CD43 abundance, Cambridge, UK). All other inhibitors and reagents data on its precise role in PMN responses are were from Sigma-Aldrich (St Louis, Missouri). scarce. CD43 is shed during PMN activation (12- Cells: Human neutrophils and endothelial cells D 15) and during adhesion and spreading (2 ,16), HUVEC were isolated as described (20,23) from ow n together with other E-selectin ligands PSGL-1 and platelet-depleted blood and umbilical vein, lo a CD44 (17,18). A soluble form of CD43, CD43s, respectively, from healthy volunteers. Neutrophils de d representing the entire extracellular domain, has were suspended in Hanks balanced solution with fro m been described in plasma and identified to Ca++ and Mg++ containing bovine serum albumin h goanl aPcMtogNl ywcoapsr octleaiinm (eGd atlog pi)n (v1o9l)v.e C sDer4i3n ep rportoetoelaysseiss 0co.1n%ce n(trHaBtioSnS +h+-aBs SnAo) . efIfne cto uorn hCaDnd4s3, ptrhoitse oBlySsAis ttp://ww w and metalloproteinases. The precise analysis of the (2). HUVECs were cultured on gelatin coated flasks .jb proteolysis mechanism was hampered by the fact (Costar Corporation, Cambrige, MA) in M199 c.o rg that CD43 soluble fragments do not transfer on supplemented with penicillin, streptomycin, L- b/ y blotting membranes (15) and required an Glutamine, Hepes (10mM), heparin (50µg/ml), g u immunoprecipitation of radiolabelled neutrophil endothelial cell growth supplement ECGS (20 est o membranes. We produced a polyclonal antibody µg/ml) and 20% fetal calf serum. 3rd- and 4th- n D against a recombinant CD43 intracellular domain, passage cells were used exclusively. CHO cells ece in order to re-assess the molecule proteolysis in the m were either transiently transfected with CD43 b e light of recent data on: i) the γ-secretase processing cDNA or stably transfected with cDNAs encoding r 1 1 of CD43 in cancer cells; ii) the description of CD43 both CD43 and core 2 beta-1,6-N- , 20 1 as a leukocyte ligand for endothelial E-selectin. We acetylglucosaminyl transferase (C2GnT), as 8 here describe for the first time, in human PMN described (24), then further transiently transfected activated by pro-inflammatory stimuli, a cathepsin to enhance protein expression. G and γ-secretase mediated processing of CD43 Neutrophil activation, flow cytometry and western with putative important functional consequences. blotting: Freshly isolated neutrophils were distributed in BSA-coated tubes and gently tumbled Experimental Procedures every 5 min during their activation at 37°C either Antibodies and reagents: FITC-anti-CD43 mAb with phorbol-myristate acetate (PMA) or with (1G10) and IgG1 were from BD Pharmingen (San TNF-α and N-Formyl-L-methionyl-L-leucyl-L- Diego, CA), anti-CD18 clone IB-4 from Ancell phenylalanine (fMLP), as described in the figure (Bayport, MN), anti-CD62E, FITC-anti-CD62L legends. The reaction was stopped by adding ice- mAbs and PE-anti-CD146 from Beckman Coulter cold PBS containing phenyl-methyl-sulfonyl- (Roissy, France). The rabbit anti-CD43cyto pAb fluoride (PMSF) 1 mM and 1,10-phenanthroline 1 was obtained as described (20). Human neutrophil mM, to prevent any further CD43 cleavage. PMN elastase (NE), cathepsin G and proteinase 3 were were centrifuged at 4°C for 30 sec at 3,800 g and purified as described (21). Purified cathepsin G was submitted to western-blotting or flow cytometry as described (20). Assays for PMN adhesion to HUVEC : 2 L Halbwachs-Mecarelli, unpublished observation 2 Static adhesion: Confluent HUVECs, sub-cultured in size to the reported soluble CD43 fragments in 48-well 0.5% gelatin coated plates, were pre- observed by immunoprecipitation of 125I-labeled treated for 4 hours with 10 ng/ml TNF-α and rinsed neutrophils (15,27): i) a cell-bound 90 kD doublet, with medium. PMN were then added (2x106 per ml resulting from clip A (Fig 1C) and complementary of M199 with 0.1% BSA) and allowed to adhere for to the reported soluble 50 kD doublet; ii) a cell- 30 min at 37°C. Non-adherent cells were removed bound CD43 C-Terminal Fragment, CD43-CTF, of by rinsing and adherent PMN and HUVEC were 25 kD, resulting from clip B and complementary to detached with 0.5 mg/ml collagenase for 15 min at the reported soluble 120 kD fragment. 37°C. They were then labeled with PE-anti-CD146, The intensity of the 90kD doublet was variable and to distinguish CD146 negative PMN from CD146 appeared to be related to the “spontaneous” positive endothelial cells. PMN were counted by activation of PMN during the isolation procedure: flow cytometry with a constant time setting. As we had observed with the soluble 50 kD doublet - Flow assay: Cell adhesion was analyzed, as (27), it was often observed on freshly isolated described (25), by real-time videomicroscopy with PMN, appeared upon incubation at cell a 10X objective, in a flow chamber containing a concentration above 106/ml without any stimuli and 0.5% gelatin-precoated glass coverslip with was significantly prevented by avoiding the lysis of confluent HUVECs, which had been stimulated for contaminating erythrocytes and by the presence of 1 4h with 0.2 ng/ml TNF-α and washed to remove the mg/ml BSA in incubation media. It is not known TNF-α. Controlled flow rates, generating a wall whether the doublet represents two very close split D shear rate of 100 s-1, were applied alternatively to sites, or a single split resulting in fragments that ow two syringes, resulting in sequential perfusion of differ by post-translational modifications such as nlo a 37°C warmed HBSS++-BSA and PMN suspension glycosylation. Kinetic studies showed that the small de d (2x106/ml), for 4 min each, over the endothelial 25 kD CD43-CTF appeared later than the 90kD fro m cells layer. This was followed by a period of doublet and was the main fragment after 30 min h wreacsohrdouint gws iwthe rwe amrmad eb ubfefetwr eaennd mviindueote m1 icarnods c2o,p aics inngc/umbla toiro nfM (FLigP. 110B-6)M. P MdidN naoctt ivmaotidoinfy b syi gTnNifFic-aαn t1l0y ttp://ww w described (26). The Histolab software (Microvision CD43, as mentioned previously (14), while priming .jb Instruments, Evry, France) allowed quantifying the with 5 ng/ml TNF-α for 30 min, followed by 30 c.o number of adherent cell visible in at least three min activation with fMLP 10-6M resulted in CD43 brg/ y fields. cleavage similar to PMA, but the 25kD CD43-CTF g u Statistical analysis: was either weakly present or not detected (Fig. 1B). est o Data were compared using a paired t test analysis. CD43 proteolysis in plasma: Western blot n D Statistical significance was defined as follows: *, p analysis confirmed that plasma inhibits CD43 ec e m < 0.05; **, p<0.01 and ***, p<0.001. proteolysis (Fig. 2A). It significantly prevented the b e decrease of CD43 full length band and the r 1 1 RESULTS appearance of CD43 fragments promoted by high , 2 0 1 Cell-associated CD43 fragments, analyzed by PMA concentrations (100 ng/ml) or, to a lesser 8 western blotting: Confirming a previous report (15), extent, by TNF-α 5ng/ml with fMLP 10-6M. This we observed that extracellular soluble CD43 inhibition, not modified by plasma heat-inactivation fragments fail to transfer onto blotting membranes, (30 min 56°C), was impressive but not complete. whatever membrane (nitrocellulose, PVDF, Although neutrophil activation in plasma resulted in positively charged PALL membranes), blotting little or no increase of the 90kD doublet, as buffer (pH 4 to 11) or anti-CD43 mAb clones (L60, compared to control cells in buffer, some 25 kD MEM59, 1G10) were used. We thus raised an anti- CD43-CTF appeared with plasma/PMA-activated CD43intra polyclonal antibody to analyze by cells (Fig. 2A). This suggests that plasma efficiently western blotting the fragments containing CD43 protects the extended extracellular domain of CD43 intracellular domains and remaining membrane- from proteolysis, while the juxtamembranous anchored after proteolysis. The antibody specificity cleavage site is less accessible to plasma inhibitors. was analyzed by western blot with CHO cell lines Sample loads of Fig 2A blots, aimed to detect the transfected with CD43 without and with Core2GnT, 25kD CD43-CTF fragment, resulted in saturating resulting in bands corresponding to the previously signals for the full length CD43 band. However, reported 115 and 130 kD glycosylated forms of serial dilutions of PMN lysates showed a significant CD43 (24) (Fig 1A). Analysis of PMA-activated decrease of full length CD43 band intensities, upon PMN (Fig.1B) confirmed previous data, since it activation in plasma with PMA and TNF/fMLP, revealed cell-associated fragments complementary respectively (Fig. 2B). 3 Role of neutrophil serine proteases in CD43 individualize, either because they translocate to the shedding: Inhibition of CD43 proteolysis may nucleus or because they are degraded by the involve plasma protease inhibitors. Since the main proteasome. PS/γ-secretase activity is thus plasma serpins are α1-proteinase inhibitor α1-PI indirectly measured by the accumulation of and α1 anti-chymotrypsin α1-CT, we analyzed the membrane CTF in the presence of γ-secretase effects of these inhibitors on CD43 proteolysis. inhibitors. The presence of three different inhibitors Western blotting analysis revealed that, while α1-PI of γ-secretase, L685,458, MG132 or compound E, does not seem to prevent any cleavage, α1-CT during neutrophil activation by PMA (Fig. 4A), inhibits the degradation of a newly appearing 70 kD significantly increased the amount of the 25kD fragment and decreases the intensity of the 25 kD CD43-CTF observed on western blots. Moreover, CD43-CTF (Fig. 3A). α1-PI is known to inhibit all in the presence of γ-secretase inhibitors, the 25kD the serine proteases released by PMN upon PMA- CD43-CTF now appeared on western blots from triggered degranulation, namely elastase, cathepsin neutrophils activated by TNF/fMLP (Fig. 4B). G and proteinase 3, PR3, with a highest affinity for Neutrophil incubation with γ-secretase inhibitors in elastase. By contrast, α1-CT inhibits cathepsin G the absence of agonist did not produce any CD43- but neither elastase nor PR3 (28). These results thus CTF fragment (data not shown). γ-secretase proposed cathepsin G as a candidate for the inhibition thus reveals the occurrence of a cleavage of the 70 kD fragment into a 25 kD cell- juxtamembranous cleavage, upon PMN activation bound CTF. Flow cytometry analysis of CD43 by TNF/fMLP, which was not detected normally, D o membrane expression, after neutrophil incubation due to CD43-CTF rapid degradation by γ-secretase. w n with exogenous purified serine proteases, showed We did not detect smaller putative intracytoplasmic loa d e no significant cleavage by PR3, while elastase and fragments resulting from this further degradation, d cathepsin G resulted in up to 80% and 40% CD43 even when cell pellets were immediately boiled in from down-regulation respectively (Fig. 3B). As shown Laemli buffer to solubilize nuclear proteins before http in figure 3A and in agreement with previous data western blot analysis (data not shown). ://w (29), western blot analysis of PMN incubated with Absence of metalloproteinase involvement in w w purified elastase resulted in a 90 kD doublet, similar CD43 shedding: RIP intramembranous proteolysis .jb c to that observed with PMA. Elastase also produced is usually triggered by matrix metalloproteinases, .org a 70 kD fragment. Interestingly, cathepsin G MMPs. We thus tested the effects of various MMP b/ y promoted a juxtamembranous cleavage, leaving a inhibitors on CD43 shedding. PMA-induced CD43 gu e s 25kD cell-associated fragment analogous to the cleavage was not inhibited by metalloproteinase t o n CD43-CTF observed after PMN activation by PMA inhibitors, such as TAPI-0, TAPI-1 and CH4644 D e (Fig. 3A). Data shown in figure 3A were obtained and CH6631, that block various MMPs, including ce m with our homemade cathepsin G preparation but MMP8 and MMP9 (data not shown) and the broad be identical results were obtained with two commercial MMP and TACE inhibitor marimastat (BB 2516) r 11 cathepsin G (data not shown). Finally, the specific (Fig.5 A,D). In parallel experiments, marimastat , 20 1 cathepsin G inhibitor 1 (10 µM) partially inhibited efficiently blocked the shedding of L-selectin 8 the appearance of the 25kD CD43-CTF resulting (CD62L) on the same cells (Fig. 5B). CD43 down- from PMA activation Fig.3C). regulation was inhibited by 1,10 phenanthroline We thus propose (Fig. 3D) that PMA-induced (Fig. 5C, D), but also by the inactive analogue 1,7 CD43 proteolysis includes: i) clips A and C, phenanthroline (Fig. 5C), confirming previous data mediated by elastase-like enzyme(s) and resulting (17) and suggesting an effect unrelated to in 90KD and 70KD membrane spanning fragments, metalloproteinase inhibition. Metalloproteinases respectively; ii) a juxtamembranous clip B, thus do not seem to be involved in PMA-induced presumably mediated by cathepsin G, producing the cleavage. 25 kD CD43-CTF. Our results also show that clip C CD43 is processed by presinilin/γ-secretase as a appears to be followed immediately by clip B and consequence of cathepsin G cleavage: We thus thus only detected if clip B is prevented by α1-CT. wondered if cathepsin G could fulfill the function of Role of presenilin/γ-secretase (PS/γ-secretase) : MMPs in the RIP processing and prepare CD43 for In cancer cells an intramembrane cleavage of CD43 γ-secretase processing. As shown in figure 6A, by PS/γ-secretase has been described (7). We thus exogenous purified cathepsin G resulted in a CD43- investigated a possible role of γ-secretase in CD43 CTF fragment, which was further degraded by γ- processing in neutrophils. In most instances, RIP- secretase. Indeed, the intensity of the CD43-CTF induced intracytoplasmic fragments are difficult to fragment, resulting from increasing doses of 4 cathepsin G, was further enhanced by PMN pre- this adhesion is β2 integrin-dependent but E- incubation with γ-secretase inhibitor compound E selectin-independent. (97.3 ± 64.6 % increase, n = 7, p = 0.007). To By contrast, PMN adhesion to TNF-activated exclude the effects of putative enzymes HUVECs under flow conditions, partially prevented contaminating our cathepsin G preparations, we by anti-E-selectin antibodies, was also significantly repeated these experiments with two commercial inhibited by 10 µg/ml Fc-CD43s, when compared cathepsin G sources and obtained the same results to human IgGs purified from an AB normal serum, (data not shown). By contrast, the intensities of as a control for the Fc portion (Fig. 8B). The CD43 fragments released by exogenous elastase inhibition level was the same with 10 µg/ml and 30 were not modified by compound E (Fig 6A). µg/ml CD43s (data not shown). In similar Finally, CD43-CTF fragments appearing upon conditions, 10 µg/ml recombinant Fc-VCAM-1 had PMN treatment by cathepsin G, in the presence of no effect on PMN adhesion (Fig. 8B). These results the γ-secretase inhibitor, disappeared completely in suggest that CD43s interferes with the E-selectin- the presence of the specific cathepsin G inhibitor I, dependent adhesion of PMN to endothelial cells but were not affected by the MMP inhibitor under flow. marimastat (Fig. 6B). This result shows that the cleavage of CD43 by purified cathepsin G, in the DISCUSSION absence of any PMN agonist, is sufficient to allow Our analysis of CD43 proteolysis, resulting from CD43 further processing by PS/γ-secretase. neutrophil activation, confirms several points D Role of presinilin/γ-secretase in the regulation of proposed previously (15). Indeed, the size of ow n neutrophil functions. γ-secretase inhibitors had no membrane-anchored C-terminal fragments, detected loa d effect on neutrophil degranulation, promoted by with our anti-CD43cyto antibody, are mostly ed PMA or TNF/fMLP, reflected by CD11b up- complementary of those previously described for fro m regulation, as measured by flow cytometry (n = 5, soluble N-terminal fragments. We demonstrate that h ttp data not shown). They had no effect either on the CD43 proteolysis includes three simultaneous or ://w oxidative burst, measured with the fluorescent sequential proteolytic steps, schematized in figure w w probe 2’-7’-dichlorofluorescein hydrodiacetate (n = 1C and 3D: Clips A and C, distant from the .jb c 5, data not shown). By contrast, all three γ-secretase membrane and a juxtamembranous clip B. .org inhibitors significantly inhibited neutrophil The double clip A, previously described as a b/ y adhesion to gelatin, promoted either by TNF/fMLP «spontaneous » cleavage (15) indeed corresponds to gu e or by PMA (Fig. 7). a protein domain very sensitive to proteolysis. This st o n Modulation of PMN adhesion to endothelial cleavage was observed upon PMN incubation D e cells by CD43s, the soluble ectodomain of CD43 without stimuli, particularly in buffers devoid of ce m released after cell activation. We tested the effect BSA and it increased with the cell concentration. b e of a recombinant chimeric Fc-CD43s molecule on This suggests that it is mediated by proteases either r 11 neutrophil adhesion to endothelial cells. To constitutively present on neutrophils, but normally , 20 1 preserve the function of the natural molecule, blocked by plasma inhibitors or albumin, or induced 8 CD43s was expressed in COS cells, which contain on PMN membrane by the isolation procedure. This the enzymatic machinery required for a cleavage was enhanced by PMA or by TNF/fMLP posttranslational glycosylation similar to and we confirm that it results in CD43 fragments neutrophils (24). We used 10 to 30 µg/ml Fc- similar to those observed with exogenous elastase CD43s, based on normal plasma CD43s (Galgp) (27,29 ). In addition, purified elastase results in a concentration, previously estimated at 10 µg/ml minor cleavage C, closer to the membrane and (19). Moreover, an home-made sandwich ELISA, leaving a 70kD membrane-spanning fragment. A using anti-CD43-coated plates and biotinylated similar 70 kD fragment was produced during PMN wheat-germ-agglutinin and standardized with the activation, but was only detected in the presence of recombinant Fc-CD43s protein, resulted in a mean α1-antichymotrypsin α1-CT, which prevents its of 7 ± 6.7 µg/ml of CD43s in normal plasma (n = 7, further degradation. These results stongly suggest data not shown). Our results show that Fc-CD43s that, during PMN activation, (an) elastase-like has no effect on PMN static adhesion to enzyme(s) cleave(s) CD43 at clip A and C sites, extracellular matrix (gelatin, fibronectin, data not distant from the cell membrane and resulting in cell shown) or to TNF-activated endothelial cells (Fig. bound 90kD doublet and 70kD fragments. The clip 8A). We confirmed with blocking antibodies that A site is more exposed and preferentially cleaved. The minor 70 kD fragment seems to be immediately 5 further processed by the α1CT-sensitive clip B the analysis of CD43 sequence (30) reveals a enzyme (as schematized in Fig. 3C) . putative cathepsin G site, Val-Pro-Phe, P -P (31), 3 1 The identification of plasma Galgp as the entire 10 aminoacids away from the transmembrane extracellular CD43 domain released in the fluid sequence. Such a cleavage would release an phase (19) makes it likely that the extracellular soluble fragment with the same C juxtamembranous clip B is involved in the terminal Phe 226 as galactoglycoprotein (Galgp), physiologic ectodomain shedding of CD43. This the soluble CD43 form described in normal plasma clip had been suggested by the down-regulation of a (19). The clip B cleavage of CD43, resulting from T305 epitope and claimed to be exclusively induced neutrophil activation, is thus likely to be mediated by PMA (15). Our anti-CD43cyto pAb allowed to by cathepsin G. clearly demonstrate, for the first time, the The lack of inhibition of PMA-induced CD43 occurrence of such juxtamembranous clip, which proteolysis by DFP, aprotinin or α1-PI, known results in a 25kD membrane inserted C-terminal inhibitors of elastase and cathepsin G, remains fragment CD43-CTF fragment. We also show that puzzling and different hypothesis have been this cleavage is not only triggered by PMA but also discussed previously (2,15). The large quantities of by pro-inflammatory stimuli, such as TNF-α proteases released by PMA-activated neutrophils priming followed by chemoattractant activation. In may overwhelm and inactivate protease inhibitors. the latter condition, the resulting CD43-CTF is only α1PI is known to be inactivated by PMN detected in the presence of γ-secretase inhibitors, as membrane-bound MMPs, MMP8 and MT6- D discussed below. Clip B also occurs during PMN MMP/MMP25 (32,33) and may not be functional at ow n apoptosis, since we previously described a 25kD sites close to PMN membrane, particularly during lo a d membrane-anchored CD43 fragment on apoptosis- cell adhesion. Moreover, a competition between e d induced membrane-derived microparticles (20). neutrophil elastase and cathepsin G for binding fro m The juxtamembranous Clip B had been proposed alpha1-PI has been described, resulting in a lack of h to result from an unknown “CD43-ase” insensitive control of cathepsin G by this inhibitor when ttp://w to serine protease or metalloproteinase inhibitors elastase is present(34). Finally the lack of effect of w w (15). We here show that, unlike clip A, which some inhibitors may be related to their accessibility .jb c occurs “spontaneously” upon incubation of isolated to the immediate membrane surrounding. Indeed, .o rg PMN, clip B requires a strong activation induced by degranulation releases endogenous proteases at the b/ y PMA or by TNF together with fMLP, while it is not immediate proximity of CD43 juxtamembranous g u e observed with TNF-α or fMLP alone. This is cleavage site. The charge or size of some but not st o reminiscent of the mobilization of azurophilic other protease inhibitors may limit their access, n D granules, which requires PMN priming together through the cell glycocalyx, to this cleavage site. ec e m with chemoattractants. Azurophilic granule serine This would also explain the partial inhibition b e proteases, which are then released, are elastase, promoted by the specific cathepsin G inhibitor I. r 1 1 cathepsin G and PR3. All are strongly cationic Following the description of a γ-secretase- , 2 0 1 enzymes, likely to spontaneously interact with such mediated RIP of CD43 in cancer cell lines (7), we 8 a negatively-charged preeminent membrane protein investigated a similar cleavage of CD43 during as CD43. Our flow cytometry data confirm that neutrophil activation. The γ-secretase component elastase and cathepsin G result in CD43 shedding, presenilin-1 has been described in neutrophil while PR3 has no significant effect (29). azurophilic granules (35). A functional presenilin/γ- Western blot analysis of CD43 proteolysis by secretase complex could then result from cathepsin G, which had not been investigated until azurophilic granule fusion with the plasma now, reveals that cathepsin G could be a good membrane. Interestingly, γ-secretase inhibitors have candidate for the juxtamembranous clip B. Indeed: recently been shown to accelerate neutrophil i) both α1-CT, which inhibits cathepsin G but apoptosis (36), suggesting that some of the γ- neither elastase nor PR3, and a specific cathepsin G secretase products may be anti-apoptotic. Upon inhibitor peptide partially inhibited PMA-induced PMN activation by PMA or by TNF-α/fMLP, we CD43 proteolysis and prevented the appearance of indeed observed a significant accumulation of the the 25 kD CD43-CTF fragment; ii) PMN treatment 25 kD CD43-CTF, when cells were pre-incubated with purified cathepsin G from three different with three different γ-secretase inhibitors. This sources resulted in a juxtamembranous cleavage, effect of γ-secretase inhibitors was also observed on producing a 25kD CD43-CTF similar to that the 25 kD CD43-CTF, resulting from PMN observed after PMN activation by PMA. Moreover, treatment by exogenous cathepsin G. The 6 juxtamembranous clip promoted by cathepsin G shedding of their extracellular region (39). Finally, could thus be sufficient to render CD43 susceptible PS/γ-secretase processing may also release CD43 to γ-secretase cleavage. This role of cathepsin G is intracellular fragments endowed with nuclear unexpected, since in most cases it is performed by localization and transcription properties. Indeed, a metalloproteinases. nuclear localization sequence has been described in A role for MMPs in CD43 shedding had been CD43 cytoplasmic tail and, in carcinoma cell lines, proposed, since PMA-induced CD43 shedding was CD43 intracellular domain was shown to inhibited by 1,10 orthophenanthroline. We tested translocate in the nucleus and cause an up here various broad metalloproteinase inhibitors, regulation of target genes (9). Gene transcription is able to inhibit MMPs and/or membrane sheddases known to be triggered in neutrophils upon such as TACE: None of them had significant effect stimulation by pro-inflammatory agonists, which on CD43 down-regulation by PMA, except for 1,10 also induce CD43 shedding. CD43 intracytoplasmic orthophenanthroline, but a similar inhibition was fragments could regulate these transcriptions, if observed with the inactive analogue 1,7 they localize in the nucleus. Genes that would be orthophenanthroline, as previously mentioned (15). differentially expressed in neutrophils in the Metalloproteinases thus do not appear to be absence or presence of γ-secretase inhibitors involved in CD43 proteolysis on PMN. deserve investigation. We addressed the question of the role of the γ- We here show that clip A is mostly prevented by secretase-mediated release of CD43 intracellular plasma, presumably due to high levels of plasma D fragments, by preventing this release with PS/γ- serpins. However, CD43 shedding has been shown ow n secretase inhibitors. Our results show that γ- to occur in vivo on circulating PMN during lo a d secretase inhibitors have no effect on neutrophil hemodialysis (42) and during PMN migration to ed degranulation or oxidative burst, but significantly inflammation sites (43). Our data suggest that the fro m inhibit neutrophil adhesion. Although this inhibition juxtamembranous clip B partially escapes plasma h may extend beyond CD43, to other putative inhibition. This cleavage was only weakly detected ttp://w membrane receptors processed by γ-secretase, these by western blotting, presumably because of the w w results suggest that neutrophil adhesion requires the immediate degradation of CD43-CTF by PS/γ- .jb c release of intracellular domains from shed secretase. Interestingly, the decrease of native .o rg transmembrane molecules. CD43, induced either by PMA or TNF/fMLP in b/ y PMN activation by inflammatory stimuli results plasma or by increasing concentrations of gu e in the shedding of half of membrane exposed exogenous cathepsin G, reached a plateau of about st o CD43. The full-length unshed CD43 molecules 50% of the initial PMN CD43 expression. This n D interact with cytoskeletal linker proteins ezrin, level is reminiscent of the 50% down-regulation of ece m radixin and moesin (ERM), are redistributed in the CD43 observed in vitro during PMN adhesion and b e uropod and participate to cell motility (37,38). The migration across activated endothelium (16) or in r 1 1 proteolysis by PS/γ-secretase of CD43-CTF vivo during PMN migration to inflamed rheumatoid , 20 1 fragments, remaining in the membrane after CD43 synovial fluids (43). This could reflect an 8 shedding, may be seen as the simple clearing of heterogeneity among CD43 molecules expressed on useless fragments by a "membrane proteasome" PMN, half of them being less exposed to cathepsin G cleavage, possibly due to steric hindrance by cis- (39). By analogy with the E-cadherin cleavage by γ- interacting membrane molecules. secretase, that dissociates catenins from the If PMN activation in whole blood only allows cytoskeleton (40), γ-secretase effect on CD43 could CD43 juxtamembranous cleavage by cathepsin G, be the release of CD43CTF-bound ERM, so that while preventing cleavages distant from the actin bundles would only interact with full-length membrane, this would release the entire CD43 molecules. Similarly, the removal of CD43- extracellular domain of CD43, CD43s. CD43s has CTF cytoplasmic domains, endowed with signaling indeed been described in human plasma and functions (4,41) may be crucial to concentrate this identified to the Galgp protein (19). We addressed signaling on CD43 molecules with entire functional the question of a possible function for CD43s, by ectodomains. Indeed, lymphocyte migration has analogy with soluble PSGL-1, which has been been shown to require both the ectodomain and the shown to have anti-inflammatory functions (49,51). intracellular domain of CD43 (4). Intramembrane Interestingly, soluble leukosialin has been reported proteolysis, such as γ-secretase-mediated RIP, may to inhibit the adhesion of colon carcinoma cells to represent a general way of switching off all signals endothelial cells (22). We postulated that CD43s, transmitted by transmembrane molecules after the acting as an E-selectin ligand (10,11), could 7 compete with cell surface selectin ligands to reduce adhesion regulatory functions. The inhibition of leukocyte rolling. Our data confirm this hypothesis, PMN adhesion to endothelium by local high since we show that CD43s inhibits the E-selectin- concentrations of CD43s may participate to the dependent adhesion of neutrophils to endothelium decrease of PMN recruitment required for the under flow conditions, while it does not interfere resolution of inflammation, as was proposed for with their integrin-mediated static adhesion. CD43s selectins ectodomain shedding (1). This has to be effects were observed at a concentration compatible added to the numerous roles of O-glycosylated with normal plasma concentrations (see results). membrane receptors in immune responses (45). Moreover, CD43s levels may be enhanced locally Neutrophil cathepsin G participates, together in inflamed microvessels, where cytokine-activated with elastase, to the shedding of various adhesion neutrophils adhere to the endothelium, a situation molecules or cytokine receptors and has been known to result in CD43 shedding(2,16). The effect shown to interfere with neutrophil migration by of CD43s is reminiscent of the reported partial several ways (reviewed in (46)). Our data inhibition of leukocyte rolling by recombinant Fc- demonstrate a new function for cathepsin G, CD43 PSGL-1, a complete inhibition being only obtained processing, which may be important for leukocyte with a mixture of PSGL-1 and L-selectin blocking migration, in view of CD43 known participation to antibodies(44). leukocyte motility and trafficking (4,37,38 ). Therefore, CD43 shedding during neutrophil activation results in fluid phase CD43s with D o w nlo a REFERENCES de d fro m h 1. Garton, K. J., Gough, P. J., and Raines, E. W. (2006) Journal of leukocyte biology 79(6), ttp 1105-1116 ://w w w 2. Nathan, C., Xie, Q. W., Halbwachs-Mecarelli, L., and Jin, W. W. (1993) The Journal of cell .jb biology 122(1), 243-256 c.o 3. Manjunath, N., Correa, M., Ardman, M., and Ardman, B. (1995) Nature 377(6549), 535-538 brg/ 4. Mody, P. D., Cannon, J. L., Bandukwala, H. S., Blaine, K. M., Schilling, A. B., Swier, K., and y g u Sperling, A. I. (2007) Blood 110(8), 2974-2982 es 5. Sikut, R., Nilsson, O., Baeckstrom, D., and Hansson, G. C. (1997) Biochemical and t on D biophysical research communications 238(2), 612-616 e c e 6. Fernandez-Rodriguez, J., Andersson, C. X., Laos, S., Baeckstrom, D., Sikut, A., Sikut, R., and m b e Hansson, G. C. (2002) Tumour Biol 23(4), 193-201 r 1 1 7. Andersson, C. X., Fernandez-Rodriguez, J., Laos, S., Baeckstrom, D., Haass, C., and Hansson, , 2 0 G. C. (2005) The Biochemical journal 387(Pt 2), 377-384 18 8. Selkoe, D. J., and Wolfe, M. S. (2007) Cell 131(2), 215-221 9. Andersson, C. X., Fernandez-Rodriguez, J., Laos, S., Sikut, R., Sikut, A., Baeckstrom, D., and Hansson, G. C. (2004) Biochemical and biophysical research communications 316(1), 12-17 10. Fuhlbrigge, R. C., King, S. L., Sackstein, R., and Kupper, T. S. (2006) Blood 107(4), 1421- 1426 11. Matsumoto, M., Atarashi, K., Umemoto, E., Furukawa, Y., Shigeta, A., Miyasaka, M., and Hirata, T. (2005) J Immunol 175(12), 8042-8050 12. Campanero, M. R., Pulido, R., Alonso, J. L., Pivel, J. P., Pimentel-Muinos, F. X., Fresno, M., and Sanchez-Madrid, F. (1991) European journal of immunology 21(12), 3045-3048 13. Kuijpers, T. W., Hoogerwerf, M., Kuijpers, K. C., Schwartz, B. R., and Harlan, J. M. (1992) J Immunol 149(3), 998-1003 14. Rieu, P., Porteu, F., Bessou, G., Lesavre, P., and Halbwachs-Mecarelli, L. (1992) European journal of immunology 22(11), 3021-3026 15. Remold-O'Donnell, E., and Parent, D. (1994) J Immunol 152(7), 3595-3605 16. Lopez, S., Seveau, S., Lesavre, P., Robinson, M. K., and Halbwachs-Mecarelli, L. (1998) Cell Adhes Commun 5(2), 151-160 17. Bazil, V., and Strominger, J. L. (1994) J Immunol 152(3), 1314-1322 8 18. Gardiner, E. E., De Luca, M., McNally, T., Michelson, A. D., Andrews, R. K., and Berndt, M. C. (2001) Blood 98(5), 1440-1447 19. Schmid, K., Mao, S. K. Y., Kimura, A., Hayashi, S., and Binette, J. P. (1980) J. Biol. Chem. 255, 3221-3226 20. Nusbaum, P., Laine, C., Bouaouina, M., Seveau, S., Cramer, E. M., Masse, J. M., Lesavre, P., and Halbwachs-Mecarelli, L. (2005) The Journal of biological chemistry 280(7), 5843-5853 21. Renesto, P., Halbwachs-Mecarelli, L., Nusbaum, P., Lesavre, P., and Chignard, M. (1994) J Immunol 152(9), 4612-4617 22. Sawada, R., Tsuboi, S., and Fukuda, M. (1994) The Journal of biological chemistry 269(2), 1425-1431 23. Jaffe, E. A., Nachman, R. L., Becker, C. G., and Minick, C. R. (1973) The Journal of clinical investigation 52(11), 2745-2756 24. Bierhuizen, M. F., Maemura, K., and Fukuda, M. (1994) The Journal of biological chemistry 269(6), 4473-4479 25. Legendre, P., Salsmann, A., Rayes, J., Trassard, O., Kieffer, N., and Baruch, D. (2006) J Thromb Haemost 4(1), 236-246 26. Radford, D. J., Luu, N. T., Hewins, P., Nash, G. B., and Savage, C. O. (2001) Arthritis and rheumatism 44(12), 2851-2861 27. Halbwachs-Mecarelli, L., Bessou, G., Lesavre, P., Renesto, P., and Chignard, M. (1996) Blood D 87(3), 1200-1202 o w 28. Travis, J., and Salvesen, G. S. (1983) Annual review of biochemistry 52, 655-709 nlo a 29. Remold-O'Donnell, E., and Parent, D. (1995) Blood 86(6), 2395-2402 de d 30. Shelley, C. S., Remold-O'Donnell, E., Davis, A. E., 3rd, Bruns, G. A., Rosen, F. S., Carroll, fro m M. C., and Whitehead, A. S. (1989) Proceedings of the National Academy of Sciences of the h United States of America 86(8), 2819-2823 ttp 31. Nakajima, K., Powers, J. C., Ashe, B. M., and Zimmerman, M. (1979) The Journal of ://w w w biological chemistry 254(10), 4027-4032 .jb 32. Owen, C. A., Hu, Z., Lopez-Otin, C., and Shapiro, S. D. (2004) J Immunol 172(12), 7791- c.o 7803 brg/ 33. Nie, J., and Pei, D. (2004) Experimental cell research 296(2), 145-150 y g u 34. Korkmaz, B., Poutrain, P., Hazouard, E., de Monte, M., Attucci, S., and Gauthier, F. L. (2005) es American journal of respiratory cell and molecular biology 32(6), 553-559 t on D 35. Mirinics, Z. K., Calafat, J., Udby, L., Lovelock, J., Kjeldsen, L., Rothermund, K., Sisodia, S. e c e S., Borregaard, N., and Corey, S. J. (2002) Blood cells, molecules & diseases 28(1), 28-38 m b e 36. Park, H. Y., Park, J. I., Baek, D. W., Lee, S. Y., Lee, M. J., Jin, J. O., Kim, J. W., Hong, Y. S., r 1 1 Lee, Y. H., and Kwak, J. Y. (2006) International immunopharmacology 6(7), 1061-1069 , 2 0 37. Sanchez-Mateos, P., Campanero, M. R., del Pozo, M. A., and Sanchez-Madrid, F. (1995) 18 Blood 86(6), 2228-2239 38. Seveau, S., Keller, H., Maxfield, F. R., Piller, F., and Halbwachs-Mecarelli, L. (2000) Blood 95(8), 2462-2470 39. Kopan, R., and Ilagan, M. X. (2004) Nature reviews 5(6), 499-504 40. Marambaud, P., Shioi, J., Serban, G., Georgakopoulos, A., Sarner, S., Nagy, V., Baki, L., Wen, P., Efthimiopoulos, S., Shao, Z., Wisniewski, T., and Robakis, N. K. (2002) The EMBO journal 21(8), 1948-1956 41. Walker, J., and Green, J. M. (1999) J Immunol 162(7), 4109-4114 42. Alvarez, V., Pulido, R., Campanero, M. R., Paraiso, V., de Landazuri, M. O., and Sanchez- Madrid, F. (1991) Kidney Int 40(5), 899-905 43. Lopez, S., Halbwachs-Mecarelli, L., Ravaud, P., Bessou, G., Dougados, M., and Porteu, F. (1995) Clinical and experimental immunology 101(1), 25-32 44. Ridger, V. C., Hellewell, P. G., and Norman, K. E. (2005) The American journal of pathology 166(3), 945-952 45. Tsuboi, S., and Fukuda, M. (2001) Bioessays 23(1), 46-53 46. Pham, C. T. (2006) Nat Rev Immunol 6(7), 541-550 9 FOOTNOTES * Authors wish to thank John Bird and Martyn Robinson, from Celltech Chiroscience Ltd, Cambridge, UK, for providing metalloproteinase inhibitors. AM was supported successively by a grant from the Association pour La Recherche sur la Polyarthrite and a grant from the Société Française d’Hématologie. We acknowledge support from the Association pour l’Utilisation du Rein Artificiel AURA, from Baxter and from Amgen. FIGURE LEGENDS Figure 1: Analysis of CD43 proteolysis by western blot with the anti-CD43cyto pAb, showing the D remaining membrane-associated CD43 fragments. A: The specificity of the anti-CD43cyto pAb was o w tested by western blotting of lysates from native CHO cells (0) and CHO transfected with CD43 nlo a without and with the Core2 GnT. Blotted membranes were analyzed with the anti-CD43cyto (right), de d while total proteins were detected with ponceau red (left). B: Kinetic analysis of PMA-induced fro proteolysis of CD43, analyzed by western blotting: Freshly isolated neutrophils (2 x 106/ml in hm HBSS++/0,1 % BSA) were incubated at 37°C with 3 ng/ml PMA and stopped at increasing incubation ttp time points (left panel) or with TNF-α 5 ng/ml for 30 min then with fMLP 10-6M for a further 30 ://ww w min-incubation, or with TNF-α or fMLP alone (right panel) and compared to cells incubated with .jb c buffer (0). Neutrophil lysates, submitted to gel electrophoresis on 7-15% gradient acrylamide, were .o rg analyzed by western blotting with the anti-CD43cyto antibody and an HRP-labeled secondary anti- b/ y rabbit IgG antibody and revealed by chemiluminescence. C: Schematic representation of soluble and g u membrane CD43 fragments. es t o n D Figure 2: CD43 proteolysis in plasma. 2 x 106/ml PMN, suspended in HBSS++-BSA or in undiluted ec e m autologous heparinated plasma, were incubated with PMA 100 ng/ml for 30 min at 37°C or with b e TNF-α 5 ng/ml for 30 min, followed by the addition of fMLP 10-6M and a further 30 min incubation. r 1 1 Cells lysates were analyzed by western blotting, as in figure 1. A: Western blot of one representative , 2 0 1 experiment and mean ± SD of the intensities of the 90 kD doublet and 25 kD scanned bands from 4 8 similar experiments. Results are expressed as percent of the maximal level observed with TNF/fMLP (90 kD doublet) or PMA (25 kD). B: Native CD43 bands shown in the blot A resulted in saturating signals. Quantitative evaluation of residual full length CD43 was then calculated from western blot analysis of serial dilutions of PMN lysates, after activation with or without plasma as above. Scanned band intensities, normalized with the scanned actin bands, are expressed as percent of the initial level of full length CD43 of PMN incubated in HBSS (n = 4). Figure 3: Role of neutrophil serine proteases in CD43 shedding. A. Left Pannel: CD43 fragments resulting from PMA activation in the presence of plasma serpins (left panel). 2 x106/ml PMN were pre-incubated for 15 min with α1-PI 500 µg/ml or α1-CT 100 µg/ml and activated for 30 min with 3 ng/ml PMA in HBSS++-BSA. Right panel: CD43 fragments resulting from the addition of exogenous purified serine proteases to resting PMN. 10 x106/ml PMN in PBS were treated with human purified elastase (NE) 15 µg/ml or cathepsin G (Cat G) 25 µg/ml for 20 min at 37°C. Cell lysates were analyzed by western blotting revealed by anti-CD43cyto pAb. B: Dose response effect of purified elastase, cathepsin G and PR3 on CD43 expression. PMN were incubated as described above with increasing concentrations of serine proteases, then labeled with FITC-anti-CD43 and analyzed by flow cytometry. C: Inhibition, by cathepsin G inhibitor 1, of the cleavage resulting in 25kD CD43- 1 0