BASIC RESEARCH www.jasn.org Increased Membrane Expression of Proteinase 3 during Neutrophil Adhesion in the Presence of Anti– Proteinase 3 Antibodies Soumeya Brachemi,* Agne`s Mambole,* Fadi Fakhouri,† Luc Mouthon,‡ Loı¨c Guillevin,‡ Philippe Lesavre,*† and Lise Halbwachs-Mecarelli* *INSERM U845 and Universite´ Paris V Rene´ Descartes, †Necker Hospital Nephrology Department, and ‡Internal Medicine Department, Cochin Hospital and Unite´ Propre de Recherche de l’Enseignement Supe´rieur (UPRES) EA 4058, Paris, France ABSTRACT We investigated membrane proteinase 3 (mPR3) expression during TNF-␣–induced adhesion of neutro- phils in the presence of anti-PR3 antibodies, a situation occurring during anti-neutrophil cytoplasmic autoantibodies (ANCA)-associated vasculitis. Three increasing levels of mPR3 expression were observed ϩ on the mPR3 neutrophil subset after stepwise cell activation. TNF-␣ activation without adhesion, TNF-␣–induced adhesion, and adhesion in the presence of anti-PR3 mAb or human anti-PR3 ANCA resulted, respectively, in a two-, seven-, and 24-fold increase of mPR3 levels. In plasma, anti-PR3 antibodies poorly recognized suspended neutrophils, whereas they bound to mPR3 on adherent cells. mPR3 upregulation was also triggered by IL-8, formyl-methionyl-leucyl-phenylalanine (fMLP), and neu- trophil adhesion to activated human umbilical vein endothelial cells. It involved ␤2 integrins and Fc␥ Ј receptor, because it was prevented by anti-CD18 antibodies and was not observed with anti-PR3 F(ab )2. Furthermore, it was specific to anti-PR3 mAb, and no mPR3 upregulation was observed with anti- myeloperoxidase or anti–HLA-ABC mAb. Newly expressed mPR3 molecules, after TNF-induced adhe- ϩ ϩ sion, were mobilized from secretory vesicles (CD35 ) and secondary granules (CD11b ). The adhesion- and antibody-dependent upregulations of mPR3 expression occurred with little azurophilic granule degranulation, no sign of apoptosis, and no further CD177 upregulation. In conclusion, this study describes an amplifying loop in polymorphonuclear neutrophil activation process, whereby ANCA are involved in the membrane expression of their own antigen during cell adhesion. This could explain the restriction of ANCA-associated vasculitis to small vessels, the main site of neutrophil adhesion. J Am Soc Nephrol 18: 2330–2339, 2007. doi: 10.1681/ASN.2006121309 Wegener granulomatosis, microscopic polyangiitis, found only after extensive cell activation.5 A high pauci-immune necrotizing glomerulonephritis, proportion of mPR3ϩ neutrophils is a risk factor and Churg-Strauss syndrome constitute a group of for the development of vasculitis, and elevated small vessel vasculitis associated with anti-neutro- mPR3 levels are associated with a higher relapse rate phil cytoplasmic autoantibodies (ANCA). The in patients with Wegener granulomatosis.6,7 In vitro pathogenesis of these vasculitides is closely linked to increased ANCA levels and polymorphonuclear neutrophils (PMN) found in the glomeruli and in- Received December 3, 2006. Accepted April 27, 2007. terstitium.1–4 Published online ahead of print. Publication date available at www.jasn.org. ANCA are directed against proteinase 3 (PR3) or myeloperoxidase (MPO), both proteinases mainly Correspondence: Dr. Lise Halbwachs-Mecarelli, INSERM U845, Hoˆpital Necker, 161 rue de Se`vres, 75015 Paris, France. Phone: stored in intracellular granules. Membrane PR3 is ϩ33-144-49-52-32; Fax: ϩ33-145-66-51-33; E-mail: mecarelli@ present on a variable proportion of resting PMN necker.fr ϩ (mPR3 ), as opposed to membrane MPO, which is Copyright © 2007 by the American Society of Nephrology 2330 ISSN : 1046-6673/1808-2330 J Am Soc Nephrol 18: 2330–2339, 2007 www.jasn.org BASIC RESEARCH ANCA-induced responses are higher in mPR3ϩ PMN than in the negative subset.8 The binding of ANCA to mPR3 may thus be an important factor in the pathogenesis of ANCA vasculitis. ANCA trigger neutrophil respiratory burst and degranulation, with the re- lease of proteolytic enzymes and reactive oxygen species.9–13 They modulate neutrophil adhesion and migration14,15 and in- duce F-actin polymerization and cell stiffening, ultimately leading to neutrophil retention in glomerular capillaries.15–17 Although the role of ANCA in vasculitis pathogenesis has been established, the accessibility of ANCA to their antigens in blood is still debated.18–20 The absence of ANCA detection on neutrophils in whole blood has been proposed to result from plasma ␣1 anti-trypsin (A1AT) binding to PR3.21 The associ- ation between A1AT-deficient phenotypes and anti-PR3–pos- itive vasculitis suggests that A1AT controls ANCA access to its antigen.22 Adherent neutrophils are surrounded by an area where pro- tease inhibitors, such as A1AT, are inactivated by neutrophil- derived metalloproteases.23–25 We hypothesized that mPR3 be- comes accessible to ANCA in the presence of plasma during neutrophil adhesion to endothelial cells, as found in vasculitis. We thus analyzed PR3 membrane expression during PMN ad- hesion induced by TNF-␣ in the presence of anti-PR3 antibod- ies, as seen in the microvascular environment of vasculitis pa- tients. RESULTS Upregulation of Membrane PR3 Expression during ␣ TNF- –Induced Adhesion Figure 1. Membrane proteinase 3 (mPR3) expression after TNF- ␣ TNF- is known to trigger adhesion-independent and -depen- ␣–induced PMN activation and adhesion. (A) Representative flow dent neutrophil responses. The addition of EDTA distin- cytometry histograms obtained with polymorphonuclear neutro- guishes the effects that are caused by TNF-␣ per se from those phils (PMN) from three healthy donors with 80% (left), 40% (mid- ϩ that result from TNF-induced integrin engagement and “out- dle), and 0% (right) of mPR3 PMN. PMN (2 ϫ 106/ml in HBSS- side-in” signaling.26 We analyzed mPR3 expression on TNF- EDTA in BSA-coated tubes) were incubated at 37°C for 45 min activated PMN from individuals with varying proportions of (Resting PMN) or, after 15 min at 37°C, for 30 min with TNF-␣ (20 ϩ ϫ 6 2ϩ PMN expressing PR3 (mPR3 ; Figure 1A). ng/ml; TNF-EDTA). PMN (2 10 /ml in HBSS -BSA) were al- ␣ lowed to settle for 15 min at 37°C in 1% gelatin–coated wells and TNF- activation without cell adhesion, in HBSS-EDTA, ␣ doubled the mPR3 level of the mPR3ϩ neutrophil subset (Fig- incubated with TNF- (20 ng/ml) for 30 min at 37°C without (TNF ϩ Adhesion) and with 2 ␮g/ml anti-PR3 mAb (TNF adhesion ϩ ure 1). Similar results were obtained with TNF-␣ activation in Ϫ anti-PR3). The anti-PR3 mAb was previously centrifuged for 20 HBSS , with or without shaking (data not shown). When neu- min at 12,000 ϫ g to remove Ig aggregates. After washing off trophils were allowed to adhere to gelatin-coated plates, upon nonadherent cells, adherent cells were released by a 20-min 2ϩ activation by TNF-␣ in HBSS , the mPR3 level was further incubation with ice-cold PBS, 1% BSA, 10 mM EDTA, and 0.1% enhanced, reaching 7 Ϯ 3-fold the resting cells level (Figure 1). sodium azide. Resting and TNF-EDTA cells were similarly incu- bated in 0.1% sodium azide before neutrophil labeling with anti- Anti-PR3 Antibody-Dependent Upregulation of PR3 mAb (solid line) or control IgG1 (dotted line) and PE-labeled Membrane PR3 Expression secondary antibody. (B) Synthesis of 22 experiments performed Ϯ ϩ We analyzed TNF-␣–activated neutrophils in the presence of as described in A (means SD of mPR3 subset mean fluores- Ͻ anti-PR3 mAb to mimic further the situation occurring in cence intensity [MFI]). ***P 0.001. ANCA-positive vasculitis. As expected, TNF-␣ induced 58 Ϯ 23% PMN adhesion, which was increased to 65 Ϯ 21% in the resulted in a stronger mPR3 upregulation, reaching 24 Ϯ 15 ϩ presence of 2 ␮g/ml anti-PR3 mAb (data not shown).14,15 Sur- times the level of resting mPR3 PMN (Figure 1). prisingly, the presence of anti-PR3 mAb throughout adhesion We excluded an in vitro artifact related to cell incubation J Am Soc Nephrol 18: 2330–2339, 2007 Neutrophil Membrane PR3 Upregulation 2331 BASIC RESEARCH www.jasn.org with mAb at 37°C, allowing a more rapid antibody binding than during the labeling performed at 4°C. The antibody con- centration was not limiting, because the mPR3 upregulation was similar with 60 ng/ml anti-PR3 mAb (n ϭ 3) instead of 2 ␮g/ml. Moreover, incubation of TNF-activated adherent PMN forupto2hat4°Cwith the mAb, to improve PR3 labeling, resulted in Ͻ10% increase of anti-PR3 binding (data not shown). This could not explain the 2.5-fold increase of mPR3 expression observed when the antibody was present during adhesion at 37°C. The mPR3 levels summarized in Figure 1B refer only to the mPR3ϩ neutrophil subset. TNF-␣ activation had no effect on the mPR3ϩ neutrophil percentage, and the mPR3Ϫ subset re- mained mostly negative during neutrophil activation and ad- hesion. It is worth noting that some mPR3 was detected on the mPR3Ϫ subpopulation as a result of TNF-induced adhesion in the presence of anti-PR3 mAb. However, this mPR3 expres- ϩ sion did not reach 3% of the level expressed by mPR3 cells in Figure 2. mPR3 expression upregulation by human anti-PR3 Ј the same conditions. ANCA. (A) Binding of anti-PR3 ANCA IgG and F(ab )2 to PMN from a donor bimodal in mPR3 expression. PMN activated by Human Anti-PR3 but not Anti-MPO ANCA Enhance TNF-␣ in EDTA, as described in Figure 1, were labeled at 4°C with mPR3 Expression during PMN Adhesion IgG (100 ␮g/ml) purified from a normal control (plain line) or from The Fc domain of ANCA or control IgG binds indistinctly to an anti-PR3 anti-neutrophil cytoplasmic autoantibodies (ANCA)- ϩ Ϫ mPR3 and mPR3 PMN subsets. As described previously,27 positive vasculitis patient (bold line, left) or with anti-PR3 ANCA Ј ␮ it leads to a high background that obscures the specific binding F(ab )2 (500 g/ml; bold line, right), followed by a secondary FITC of anti–PR3-ANCA IgG (Figure 2A, left).