Vasoactive Kinin Kallikrein-Independent Release Of

Neutrophil-Derived Proteinase 3 Induces Kallikrein-Independent Release of a Novel Vasoactive Kinin

This information is current as Robin Kahn, Thomas Hellmark, L. M. Fredrik of September 23, 2021. Leeb-Lundberg, Nasrin Akbari, Mihail Todiras, Tor Olofsson, Jörgen Wieslander, Anders Christensson, Kerstin Westman, Michael Bader, Werner Müller-Esterl and Diana Karpman

J Immunol 2009; 182:7906-7915; ; Downloaded from doi: 10.4049/jimmunol.0803624 http://www.jimmunol.org/content/182/12/7906

References This article cites 44 articles, 14 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/182/12/7906.full#ref-list-1

Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists by guest on September 23, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

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

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2009 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Neutrophil-Derived Proteinase 3 Induces Kallikrein-Independent Release of a Novel Vasoactive Kinin1,2

Robin Kahn,* Thomas Hellmark,† L. M. Fredrik Leeb-Lundberg,‡ Nasrin Akbari,* Mihail Todiras,§ Tor Olofsson,¶ Jo¨rgen Wieslander,ʈ Anders Christensson,# Kerstin Westman,† Michael Bader,§ Werner Mu¨ller-Esterl,** and Diana Karpman3*

The kinin-forming pathway is activated on endothelial cells and neutrophils when high-molecular weight kininogen (HK) is cleaved by plasma kallikrein liberating bradykinin, a potent mediator of inflammation. Kinins are released during inflammatory conditions such as vasculitis, associated with neutrophil influx around blood vessels. Some patients with vasculitis have elevated plasma levels of neutrophil-derived proteinase 3 (PR3) and anti-PR3 Abs. This study investigated if neutrophil-derived PR3 could induce activation of the kinin pathway. PR3 incubated with HK, or a synthetic peptide derived from HK, induced breakdown and Downloaded from release of a novel tridecapeptide termed PR3-kinin, NH2-MKRPPGFSPFRSS-COOH, consisting of bradykinin with two addi- ␣ tional amino acids on each terminus. The reaction was specific and inhibited by anti-PR3 and 1-antitrypsin. Recombinant wild-type PR3 incubated with HK induced HK breakdown, whereas mutated PR3, lacking enzymatic activity, did not. PR3-kinin bound to and activated human kinin B1 receptors, but did not bind to B2 receptors, expressed by transfected HEK293 cells in vitro. In human plasma PR3-kinin was further processed to the B2 receptor agonist bradykinin. PR3-kinin exerted a hypotensive effect in vivo through both B1 and B2 receptors as demonstrated using wild-type and B1 overexpressing rats as well as wild-type and B2 receptor knockout http://www.jimmunol.org/ mice. Neutrophil extracts from vasculitis patients and healthy controls contained comparable amounts of PR3 and induced HK proteolysis, an effect that was abolished when PR3 was immunoadsorbed. Neutrophil-derived PR3 can proteolyze HK and liberate PR3- kinin, thereby initiating kallikrein-independent activation of the kinin pathway. The Journal of Immunology, 2009, 182: 7906–7915.

n vivo activation of the kinin pathway occurs when the cir- results in the generation of active plasma kallikrein that in turn

culating plasma prekallikrein-high-molecular weight kinino- cleaves HK and liberates bradykinin (1). Bradykinin binds to B2 I gen (HK)4 complex binds to receptors present on endothelial receptors constitutively expressed on endothelial cells and is de- cells and neutrophils. Proteolytic cleavage of plasma prekallikrein graded by carboxypeptidases to des-Arg9-bradykinin, which binds by guest on September 23, 2021 to B1 receptors induced during inflammation (2, 3). Binding of bradykinin to its receptor results in capillary leakage, inflamma- *Department of Pediatrics, Clinical Sciences Lund, †Department of Nephrology, Clin- tion, pain, and local reduction of blood pressure. Bradykinin also ical Sciences Lund, ‡Unit of Drug Target Discovery, Division of Cellular and Mo- lecular Pharmacology, Department of Experimental Medical Science, Lund Univer- stimulates intimal hypertrophy and proliferation of smooth muscle sity, Lund, Sweden; §Max-Delbru¨ck-Center for Molecular Medicine, Berlin-Buch, cells during endothelial damage (1). The kinin system is activated Germany; ¶Division of Hematology and Transfusion Medicine, Department of Lab- oratory Medicine, Lund University, Lund, Sweden; ʈWieslab, Lund, Sweden; #De- during several conditions associated with systemic inflammation partment of Nephrology and Transplantation, Clinical Sciences Malmo¨, Lund Uni- such as vasculitis (4), sepsis (1), and colitis (5). versity, Lund, Sweden; and **Institute of Biochemistry II, University Hospital, Neutrophil extravasation into inflamed tissues with degranula- Frankfurt, Germany tion and release of proteolytic enzymes, such as proteinase 3 (PR3) Received for publication October 31, 2008. Accepted for publication April 9, 2009. (6), is one of the major features of systemic inflammatory condi- The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance tions such as vasculitides. PR3 is encoded by a gene located on with 18 U.S.C. Section 1734 solely to indicate this fact. chromosome 19p13.3 consisting of five exons. The catalytic site of 1 This work was presented in part in poster form at the 13th Congress of the Inter- PR3 consists of His44, Asp91, and Ser176, located on exons two, national Pediatric Nephrology Association, August 29 to September 2, 2004, Ad- three and five, respectively. In resting neutrophils, PR3 is stored in elaide, Australia, and the 2nd International Conference on Exploring the Future of Vascular and Inflammatory Mediators, Kinin 2007, May 30 to June 2, 2007, Berlin, the azurophilic and specific granules as well as secretory vesicles, Germany. or it is expressed on the plasma membrane. The acidic environ- 2 This study was supported by grants from the Swedish Research Council (K2007- ment of the azurophilic granules keeps PR3 in a conformationally 64X-14008-07-3), the fund for Renal Research, Crown Princess Lovisa’s Society for Child Care, the Sven Jerring Foundation, Konung Gustaf V:s 80-årsfond, Fanny Ek- inactive form. Upon cell activation degranulation and translocation dahl’s Foundation (all to D.K.) and Queen Silvia’s Jubilee fond (to R.K.) and the to an environment with neutral pH enables the protease to become Swedish Research Council (grant 15057; to L.M.F.L.L.). Diana Karpman is the re- cipient of a clinical-experimental research fellowship from the Royal Swedish Acad- enzymatically active (reviewed in Ref. 6). When neutrophils are emy of Sciences. activated, membrane-bound PR3 is up-regulated (7) and a subset 3 Address correspondence and reprint requests to Dr. Diana Karpman, Department of of patients with vasculitis has more PR3 on their neutrophil mem- Pediatrics, Lund University, 22185 Lund, Sweden. E-mail address: Diana.Karpman@ branes as well as higher levels of PR3 in plasma than do controls med.lu.se (8). PR3 degrades extracellular matrix proteins, kills microbes (6), 4 Abbreviations used in this paper: HK, high-molecular weight kininogen; PR3, pro- ␣ ␣ teinase 3; 1-AT, 1-antitrypsin; ANCA, anti-neutrophil cytoplasmic Abs; KLKs, and has the ability to cleave and inactivate C1 inhibitor (9). PR3 kallikreins; EACA, ␧-amino-n-caproic acid; GEMSA, 2-guanidinoethylmercaptosuc- binds to endothelial cells (10) and induces apoptosis (11) and IL-8 cinic acid; ESI, electrospray ionization; PI, phosphoinositide. production and secretion from these cells (6). The physiological ␣ ␣ Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00 inhibitor of PR3 in plasma is 1-antitrypsin ( 1-AT) (12). www.jimmunol.org/cgi/doi/10.4049/jimmunol.0803624 The Journal of Immunology 7907

In this study we investigated vasculitis as a model for systemic Bradykinin ELISA inflammation with activation of the kinin pathway (4). Vasculitis is Samples were analyzed for bradykinin release using the MARKIT-M bra- a systemic condition manifested by inflammation in and around dykinin-ELISA kit (Dainippon Pharmaceutical) as per the manufacturer vessel walls with neutrophil infiltration accompanied by perturbed and as described (4). To determine the optimal time for incubation of PR3 vessel patency and secondary tissue damage. Many different or- with purified HK, purified HK (5 ␮g/ml) was incubated with PR3 (1 and ␮ gans may be involved, such as the kidneys, lungs, skin, intestines, 10 g/ml) for 1, 5, 10, 20, 30, 45, 60, and 90 min. Liberation of kinins peaked at 20 min and declined toward 90 min. The length of incubation of and joints. Patients exhibit various symptoms such as renal dys- PR3 with purified HK was therefore set at 20 min. function, purpura, abdominal pain, and joint pain and swelling (13–15). Severe vasculitis is often associated with the presence of Mass spectrometry anti-neutrophil cytoplasmic Abs (ANCA) directed toward granular For analysis by MALDI, HK (400 ␮g/ml, final concentration) was incu- components of neutrophils, predominantly PR3 or myeloperoxi- bated with PR3 (100 ␮g/ml, final concentration) or with plasma kallikrein ␮ dase (6). (326 g/ml, final concentration) at 37°C for 20 min. Trichloroacetic acid (18.5 mg/ml, final concentration; Dainippon Pharmaceutical) was added to The components of the activated kinin system may contribute to the samples, after which samples were centrifuged at 1300 ϫ g for 10 min the inflammatory process during vasculitis (4). As neutrophil in- and analyzed directly. An aliquot of 0.5 ␮l was applied to a MALDI target filtrates are a cardinal pathological feature of vasculitis, and acti- slide surface and allowed to dry. One-half microliter of matrix solution (2.5 vated neutrophils express and secrete PR3, the present study aimed mg/ml of ␣-cyano-4-hydroxy-cinnamic acid in 50% acetonitrile containing ␮ to investigate if PR3 from neutrophils could cleave HK and liber- 0.05% trifluoroacetic acid) and another 0.5 l of the sample were added and dried before being used for analysis on a MALDI-TOF mass spec- ate bradykinin or bradykinin-like peptides in a kallikrein-indepen- trometer (MALDI micro MXT; Waters, Manchester, U.K.). dent manner, establishing a novel kinin-forming pathway that may In separate experiments synthetic PR3-kinin (position 361–374 in HK, Downloaded from operate during systemic inflammation. MKRPPGFSPFRSS; final concentration, 0.1 mg/ml; MedProbe) was incubated with platelet-poor plasma, plasma kallikrein (final concentration, 163 ␮g/ml), or PBS for 30 s or 1, 5, or 10 min. Additionally, PR3-kinin was Materials and Methods incubated for 10 min with platelet-poor plasma, which was preincubated ␧ PR3-induced HK proteolysis for 20 min with carboxypeptidase inhibitors: -amino-n-caproic acid (EACA; Kebo Lab), which inhibits carboxypeptidase B; 2-guanidinoeth- To investigate if PR3 could proteolyze HK, purified HK (100 ␮g/ml, cor- ylmercaptosuccinic acid (GEMSA; Calbiochem), which inhibits car- http://www.jimmunol.org/ responding to the plasma concentration (1); Enzyme Research Laborato- boxypeptidase B-like processing enzyme; and Plummer’s inhibitor (Cal- ries) was incubated for 20 min at 37°C with PR3 (Wieslab) at a final biochem), which inhibits carboxypeptidase N. PR3 was also separately concentration of 1 or 10 ␮g/ml (5ϫ and 50ϫ the concentration in plasma incubated with platelet-poor plasma preincubated for 60 min with Com- from vasculitis patients (16, 17)) or with active plasma kallikrein (Enzyme plete Mini (Roche), a protease inhibitor cocktail that inhibits serine, cys- Research Laboratories) at a final concentration of 16.3 ␮g/ml, as a positive teine, and metalloproteases, as well as calpains. The samples were analyzed control. Recombinant PR3 (10 ␮g/ml, final concentration), wild-type, or by MALDI as above. enzymatically inactivate PR3 (18, 19) were incubated with HK (100 ␮g/ml, To analyze the amino acid sequence of released kinins and rule out the final concentration) for 20 min at 37°C. Specificity was tested by prein- possibility of contaminating proteases (such as plasma kallikrein) in puri- cubating PR3 with polyclonal rabbit anti-human PR3 (Wieslab) at 67 fied HK, a peptide of 28 aa residues covering the sequence of the human ␮ g/ml (1:1 molar ratio) for 30 min at 37°C before adding HK. The anti- HK domain 4 (D4; position 354–381) was synthesized (MedProbe): NH2-

PR3 Ab did not cross-react with plasma kallikrein (0.5 ␮g), tissue kal- PLGMISLMKRPPGFSPFRSSRIGEIKEE-COOH, referred to as synthetic by guest on September 23, 2021 likrein (KLK1; Abnova), or kallikreins (KLKs) 4, 5, and 8 (a gift from HK D4. This synthetic HK D4 consists of bradykinin (underlined) with an Maria Brattsand, Department of Public Health and Clinical Medicine, additional 9 and 10 residues at the N terminus and C terminus, respectively, Umeå University, Sweden), all of which were run on a 10% SDS-PAGE and the amino acid sequence matches parts of the D4 domain identical in gel (0.5 ␮g of kallikrein per row) and electroblotted. The immunoblot was both HK and low-molecular weight kininogen. Synthetic HK D4 (at a final compared with a protein gel exposed to SilverSnap stain (Thermo concentration of 9 mg/ml) was incubated with PR3 at a final concentration Scientific). of 13 or 130 ␮g/ml or with plasma kallikrein (the positive control) at a final PR3 (1 and 10 ␮g/ml, final concentration) was also incubated with concentration of 163 ␮g/ml at 37°C for 20 min and analyzed by MALDI ␣ 1-AT (1.8 mg/ml, final concentration, corresponding to the known plasma (as above) or by electrospray ionization (ESI) with a QToF Ultima API concentration (20), gift from Dr. Sabina Janciauskiene, Lund University, (Waters) coupled to a CapLC HPLC. The digests were separated on re- Sweden), known to rapidly inhibit PR3 activity in plasma (12), for 5 min versed-phase analytical column (Atlantis; C18, 75 ϫ 150 mm, 3 mm, 100 at 37°C and then added to HK (100 ␮g/ml, final concentration) for 20 min Å; Waters) using a 60 min linear gradient. Dynamic data acquisition scan- at 37°C. PR3 (10 ␮g/ml, final concentration) was incubated with normal ning was used over the mass range m/z from 400 to 1600 for MS (MALDI) platelet-poor plasma (obtained as in Ref. 4), concentration of HK Ϸ80 and from 50 to 1800 for MS/MS (ESI). Only spectra from ions with charge ␮g/ml (1) for 90 min at 37°C. To investigate if PR3 liberates kinins from state 2 and 3 were acquired. Database searches were done using the known HK (by ELISA), purified HK (5 ␮g/ml) was incubated for 20 min with PR3 sequence of HK. at a final concentration of 1 or 10 ␮g/ml at 37°C or with plasma kallikrein at a final concentration of 16.3 ␮g/ml, as the positive control. Samples were Competition binding assays Ϫ Ϫ frozen at 20°C (for immunoblotting) and at 80°C (for assay by HEK293 cells (American Type Culture Collection), stably transfected with ELISA). the human kinin B1 receptor and B2 receptor, were cultured as described (23). Radioligand binding with 1 nM [3H]des-Arg10-kallidin (77.5 Ci/ Detection of HK proteolysis by immunoblotting mmol) and 1 nM [3H]bradykinin (90.0 Ci/mmol) (PerkinElmer) was conducted as previously described (3). Binding assays were conducted on ice Proteolytic processing of HK was assayed by immunoblotting as previ- in duplicates, and nonspecific binding was defined as the amount of radio- ously described (4). Samples containing HK were diluted 1/10 and samples labeled ligand bound in the presence of 1 ␮M unlabeled des-Arg10-kallidin containing plasma were diluted 1/100. The samples were run on a SDS- or bradykinin. The IC50 was determined from the dose-dependent displace- 10% PAGE gel, electroblotted for 2 h, and membranes were incubated with ment of 1 nM [3H]des-Arg10-kallidin (for the B receptor) or 1 nM [3H]bra- a primary Ab against HK. Three different primary Abs (4) were used: 1 dykinin (for the B2 receptor) by PR3-kinin. Binding data were analyzed antiserum AS88 (from sheep) against purified HK (1/5000) (21), mAb using Prism (GraphPad Software). HKH8 (from mouse) directed to domain D2 (1/1000) (22), and mAb HKL22 directed to domain D6H (1/500) (22). Appropriate secondary Abs Kinin receptor activation quantified by phosphoinositide (PI) were applied, including goat anti-mouse/HRP (1/3000), goat anti-rabbit/ hydrolysis HRP (1/3000; both Abs from Dako), and donkey anti-sheep/HRP (1/3000; ICN Biomedicals). The positive controls were either HK incubated with HEK293 cells were seeded in 6-well dishes and labeled with 1 ␮Ci/ml plasma kallikrein or platelet-poor plasma incubated with dextran sulfate [3H]myo-inositol (PerkinElmer) in DMEM/10% FBS for 20–24 h at 37°C. ␮ (75 mol/L; Sigma-Aldrich) in the presence of ZnCl2 (2 mmol/L; Sigma- The cells were then washed four times in DMEM, and further incubated in Aldrich) (4). The negative controls were HK incubated with PBS (Medi- DMEM for1hat37°C. This was followed by incubation in the presence cago) or plasma with PBS. The signal was detected by chemiluminescence. or absence of PR3-kinin in DMEM supplemented with 50 mM LiCl for 30 7908 PROTEINASE 3 RELEASES A NOVEL VASOACTIVE KININ

Table I. Clinical and laboratory features of patients at time of sampling

Proteolysis Bradykinin Patient Age at of HK in Level in Plasma No. Sex Samplinga Diagnosis ANCA Plasma (pg/ml) Symptoms, Laboratory Findings, and Pathology

1 F 76 WG PR3 ϩ 1908 Gingivitis, rhinitis, hematuria 2 F 61 WG MPO ϩ 1182 Polycondritis, cough, rhinitis, hematuria 3 M 80 MPO-ANCA associated MPO ϩ 727 Rash, thrombosis, and necrosis in renal biopsy vasculitis 4 F 11 PR3-ANCA associated PR3 ϩ 340 Rash, arthralgia vasculitis 5 F 14 PR3-ANCA associated PR3 ϩ 472 Rash, arthralgia, elevated creatinine vasculitis 6 F 12 WG PR3 and MPO ϩ 1488 Fever, arthralgia, elevated creatinine, glomerulonephritis in renal biopsy 7 M 16 HSP ϩ 256 Rash, arthralgia, gastrointestinal symptoms, glomerulonephritis in renal biopsy 8 F 18 HSP ϩ 299 Elevated creatinine, glomerulonephritis in renal biopsy 9 F 11 HSP ϩ 156 Rash, fever, arthralgia 10 F 10 HSP ϩ 81 Rash, arthralgia, gastrointestinal symptoms, elevated creatinine Downloaded from

a Age at sampling is the same as age at onset of disease. F, female; M, male; ϩ, HK proteolysis visible in plasma samples assayed by immunoblotting (4); WG, Wegener’s granulomatosis; MPO, myeloperoxidase; HSP, Henoch Schönlein purpura.

min at 37°C. The cells were then lysed with 0.1 M formic acid for 20 min described (4). The clinical and laboratory aspects of all patients are sum- at 4°C and centrifuged at 16,100 ϫ g for 5 min at 4°C. The supernatants marized in Table I. Informed consent was obtained from all participants were added to anion exchange columns, which were washed twice with a and the study was approved by the Ethics Committee of the Medical Fac- http://www.jimmunol.org/ low salt solution (60 mM ammonium formate, 5 mM sodium borate). Ino- ulty of Lund University. sitol phosphates were then eluted with a high salt solution (1 M ammonium formate, 0.1 M formic acid) and radioactivity was quantiﬁed in a Beckman LS6000 liquid scintillation counter (Beckman Coulter). Blood samples Venous blood from patients and controls were collected in 2.7-ml plastic In vivo experiments using wild-type and B2 receptor knockout tubes (BD Diagnostics) containing 0.129 M sodium citrate. Plasma was mice obtained by centrifugation as previously described (4). Whole blood for neutrophil isolation was obtained in 15-ml plastic tubes (Sarstedt) with Adult (12 wk old) male C57BL/6 wild-type and kinin B receptor knockout 2 sodium heparin (150 USP, Venoject; Terumo Europe).

mice maintained on the same background (24) were used as previously by guest on September 23, 2021 described (25). Blood pressure and heart rate were recorded with a trans- ducer (model 1050.1; AD Instruments) that was connected to a computer Evidence for HK proteolysis and bradykinin release in patients system for data acquisition and analysis (Chart 5.5.5 PowerLab; AD In- and controls struments) in freely moving mice. Bradykinin (Bachem) and PR3-kinin were given as a bolus dose (1 ␮l/10 g of body weight) in cumulative Criteria for inclusion of patients in this study were the presence of HK concentrations (0.1–100 ␮g/kg) at 45- to 60-min intervals. proteolysis and bradykinin release in plasma samples. HK proteolysis and bradykinin release were analyzed in patients and controls as previously In vivo experiments using wild-type and B1 receptor described (4) and are presented in Table I. Plasma samples were available overexpressing rats from 10 vasculitis patients (pediatric and adult) and 12 adult controls. Samples from patients were taken at the onset of symptoms. All patient Twelve-week-old male rats (350–400 g) from colonies of Hannover samples exhibited HK proteolysis and bradykinin release in plasma sam- Sprague-Dawley rats and heterozygous TGR-Tie2B1 transgenic rats on the ples with a median of 406 pg/ml (range, 81–1908 pg/ml) (Table I), whereas Hannover Sprague-Dawley background were maintained at the Max-Del- the controls did not exhibit any HK proteolysis and their bradykinin levels bru¨ck-Center. The generation of TGR-Tie2B1 rats overexpressing the B1 were considerably lower, with a median of 86 pg/ml (range, 0–425 pg/ml). receptor exclusively in the endothelium and the experimental procedures were previously described (26). Substances were given in 1 ␮l per 100 g of body weight at the following doses: 0.1 ␮g/kg des-Arg10-kallidin Neutrophil purification ␮ (Bachem); 1 and 10 g/kg bradykinin and PR3-kinin. Icatibant (HOE 140, Neutrophils were purified from five patients and five controls (of which ␮ 100 g/kg; Sigma-Aldrich) was administered i.v. 10 min before adminis- four controls also contributed plasma samples). Neutrophils were isolated tration of bradykinin or PR3-kinin. Blood pressure was measured as pre- from whole blood with Polymorphprep (Axis-Shield) or NIM (Cardinal viously described (26). Associates) as per the manufacturers’ instructions. The neutrophils were All experimental procedures in mice and rats were performed in accor- resuspended at a concentration of 1 ϫ 106 cells/ml in a buffer containing dance with the guidelines for the humane use of laboratory animals as 5.6 mM D -glucose, 127 mM NaCl, 5.4 mM KCl, 1.2 mM KH PO , 0.8 approved by the local ethical committee of Max-Delbru¨ck-Center for Mo- 2 4 mM MgSO4, 10 mM HEPES, and 1.8 mM CaCl2 (28). The cells were lecular Medicine, Berlin. identified as leukocytes by flow cytometry using anti-CD45:FITC, and the neutrophil population was identified using three separate Abs, anti-CD11b: Subjects RPE, anti-CD18:FITC, and anti-CD16:FITC. The lymphocyte population Healthy adult controls (n ϭ 13, 8 women, 5 men) and patients with vas- was identified using anti-CD3:FITC, anti-CD19:FITC, and a platelet pop- culitis (n ϭ 10, 3 adults, 7 children) participated in the study. Vasculitis ulation was ruled out using anti-CD41:RPE (all Abs from Dako; 5 ␮lofAb was defined according to Chapel Hill nomenclature (27). Neither the pa- to 50 ␮l of cells). The cells isolated consisted only of leukocytes, of which tients nor the controls were receiving any medications at the time of sam- at least 90% were neutrophils. pling, and all patient samples were taken at the onset of vasculitis symptoms. Patients were admitted to the Department of Nephrology, University Preparation of neutrophil extracts Hospital, Malmo¨, Sweden, between September and October 2002 (n ϭ 3, patients 1–3) or the Department of Pediatrics, Section of Pediatric Ne- The neutrophils were incubated for 30 min with or without 0.05% Triton phrology, University Hospital, Lund, Sweden, between June 1999 and Sep- X-100 (Merck) on ice to induce cell lysis and release of the cell content to tember 2001 (n ϭ 7, patients 4–10). Patients 4–10 have been previously the medium. The cell extracts were then centrifuged at 14,000 ϫ g for 1 The Journal of Immunology 7909

FIGURE 1. Breakdown of HK in plasma and in purified form by PR3 and plasma kallikrein. Lanes 1–6, Immunoblots detected with AS88, a polyclonal Ab recognizing all of the domains of HK and its breakdown products. Lane 1, Control plasma, showing two bands at 116 and 63–66 kDa corresponding to HK and its H chain and low-molecular weight kininogen (4). Lane 2, Normal plasma with dextran sulfate inducing kinin pathway activation showing three bands at 63–66, 58, and 45 kDa, corresponding to H chain, L chain, and degradation of the L chain. Lane 3, Representative plasma from an adult vasculitis patient, showing HK proteolysis. Lane 4, Purified HK (100 ␮g/ml). Lane 5, Purified HK incubated with plasma kallikrein (16.3 ␮g/ml) inducing breakdown of HK, showing bands at 63–66 and 45 kDa. Lane 6, Purified HK with PR3 (10 ␮g/ ml) inducing breakdown of HK, showing four Downloaded from bands, corresponding to an unidentified breakdown product (70 kDa), H chain (63 kDa), L chain (58 kDa), and a degradation product of the L chain (45 kDa). Lane 7, Detected with HKH 8, mAb recognizing domain 2 (H chain): purified HK incubated with PR3 (10 ␮g/ml) inducing

breakdown of HK, showing one band at 63 kDa http://www.jimmunol.org/ corresponding to the H chain. Lane 8, Detected with HKL 22, mAb recognizing domain 6 (L chain and degradation of L chain): puriﬁed HK incubated with PR3 (10 ␮g/ml) inducing breakdown of HK, showing three bands at 70, 58, and 45 kDa corresponding to an unidentiﬁed breakdown product, L chain, and degradation of L chain. D1–D6, domains 1–6 of high-molecular weight kininogen. by guest on September 23, 2021

min. These extracts contain cell lysates and membranes. No IgG was de- Wilcoxon signed ranks test was used for comparison of differences in kinin tected in the neutrophil extracts from patients or controls by immunoblot- release from HK by the neutrophil extracts between the four experimental ting using rabbit anti-human IgG-HRP Ab (Dako). The supernatants were groups. A p value of Յ0.05 was considered significant. SPSS version 14 transferred into plastic tubes and frozen at Ϫ80°C until analyzed. was used for the statistical analyses. Immunoadsorption of PR3 Results A CNBr-activated Sepharose 4B gel (Amersham Biosciences) was coupled PR3 proteolysis of HK resembles the physiological breakdown with the mAb 4A3 against PR3 (29) (0.4 mg/ml, 1.2 mg to 0.5 g Sepharose pattern gel). Neutrophil extracts from patients and controls were incubated individually with the Sepharose gel for 3 h at room temperature and the flow- HK consists of six domains representing the H chain, the brady- through was collected. The 4A3 Ab does not react with other neutrophil kinin sequence, and the L chain (Fig. 1) (1). In vitro degradation proteases, such as elastase and cathepsin G (29, 30), or with plasma kal- of HK in plasma by dextran sulfate results in three fragments vis- likrein, tissue kallikrein (KLK1), or KLKs 4, 5, and 8 (method as described ible by immunoblotting using the polyclonal AS88 Ab, which rec- above for the rabbit anti-human PR3 Ab). Levels of PR3 were measured in neutrophil extracts by ELISA, as previously described (31). The efficiency ognizes all domains of HK including the H chain (63 kDa, domains of immunoadsorption was determined in this manner. 1–3), the L chain (58 kDa, domains 5 and 6), and a breakdown product of the L chain (45 kDa, domain 6; Fig. 1, lanes 1 and 2) Neutrophil extract-induced HK proteolysis (4). These three fragments resemble the physiological breakdown Four variants of neutrophil extracts were tested: Triton X-100-treated neu- pattern in plasma seen during kinin system activation in vasculitis trophil extract, Triton X-100-treated neutrophil extract in which PR3 was (Fig. 1, lane 3) (4). Plasma kallikrein degrades purified HK (1, 32) immunoadsorbed, untreated neutrophil extract, and untreated neutrophil into two fragments (63 and 45 kDa; Fig. 1, lane 5). When HK was extract in which PR3 was immunoadsorbed. The neutrophil extracts were added to purified HK (5 ␮g/ml for ELISA and 100 ␮g/ml for immuno- incubated with PR3, three bands were detected (Fig. 1, lane 6), blotting) at volume proportions of 1:9 (neutrophil extract/HK). As a con- corresponding to the bands seen in plasma. mAbs against domains trol, the neutrophil buffer without neutrophils was combined with purified 2 or 6 identified these bands as the H chain, L chain, and the HK. The samples were incubated for 20 min at 37°C and assayed as above. breakdown product of the L chain (Fig. 1, lanes 7 and 8) (4). A ϳ Statistics fourth band at 70–75 kDa was also identified, which may rep- resent a breakdown product of HK. The Mann-Whitney U test was used for comparing the difference between These results indicate that the PR3-induced proteolysis pattern the levels of kinins released from HK by two concentrations of PR3 and the negative control and for comparison of differences in kinin levels between of HK resembles HK breakdown in plasma and is similar, but not neutrophil extracts from the patients and controls and of differences in PR3 identical, to the pattern produced by the known physiological pro- levels in the neutrophil extracts between the patients and the controls. The cessing enzyme, plasma kallikrein. 7910 PROTEINASE 3 RELEASES A NOVEL VASOACTIVE KININ Downloaded from http://www.jimmunol.org/ by guest on September 23, 2021

FIGURE 2. PR3 induced HK degradation. A, PR3 induces breakdown of HK. Lane 1, Immunoblot showing purified HK (100 ␮g/ml) incubated with plasma kallikrein (16.3 ␮g/ml), inducing total HK proteolysis. Lane 2, Purified HK (100 ␮g/ml). Lane 3, HK incubated with PR3 (10 ␮g/ml), inducing almost total breakdown of HK. Lane 4, HK incubated with PR3 (1 ␮g/ml), inducing partial breakdown of HK. Samples were run on the same gel. The experiment was repeated five times with reproducible results. One representative experiment is shown. The ELISA results showing bradykinin levels (in pg/ml, median (range)) are depicted under the immunoblot figure. Purified HK (5 ␮g/ml) was incubated with PR3 (1 and 10 ␮g/ml). ELISA experiment was repeated eight times. Differences in bradykinin release after incubation with PR3 were significant when comparing the two different PR3 concentrations individually with the negative control as well as when comparing the two different PR3 concentrations with each other (p ϭ 0.001). B, HK proteolysis of PR3 was dependent on enzymatic activity. Lane 1, Purified HK (100 ␮g/ml). Lane 2, HK incubated with PR3 (10 ␮g/ml), inducing breakdown of HK. Lane 3, HK incubated with recombinant PR3 (10 ␮g/ml), inducing partial breakdown of HK. Lane 4, HK incubated with recombinant mutated PR3 (10 ␮g/ml), resulting in no breakdown of HK. The experiments were repeated three times and one representative gel is shown. C, Specificity of PR3-induced proteolysis. Lane 1, Purified HK (100 ␮g/ml). Lane 2, HK incubated with PR3 (10 ␮g/ml), inducing breakdown of HK. Lane 3, PR3 preincubated with an anti-PR3 Ab (67 ␮g/ml) and then incubated with HK. The Ab could partially inhibit HK proteolysis. The experiment was repeated three times and the lanes were ␣ ␮ ␮ run on the same gel. D, 1-Antitrypsin inhibits PR3-induced breakdown of HK. Lane 1, Purified HK (100 g/ml). Lane 2, PR3 (10 g/ml) incubated with ␣ HK, inducing total breakdown of HK. Lane 3, PR3 incubated with 1-antitrypsin (1.8 mg/ml) before combining with HK resulted in total inhibition of PR3-induced breakdown of HK. The samples were run on the same gel. The experiment was repeated four times.

PR3 proteolyzes HK in a dose-dependent and specific manner HK proteolysis (Fig. 2D, lanes 2 and 3). Purified PR3 was added Incubation of purified HK (Fig. 2A, lane 2) with PR3 resulted in a to normal plasma to investigate whether the proteolytic activity of dose-dependent breakdown of HK (Fig. 2A, lanes 3 and 4) and PR3 was functional in normal plasma, and no proteolysis of HK release of kinins (Fig. 2A, lanes 3 and 4). Incubation of HK with was detected (data not shown). The results indicate that PR3 pro- ␣ recombinant wild-type PR3 resulted in partial degradation of HK teolysis of HK is inhibited by 1-AT, and this most probably ac- (Fig. 2B, lane 3), whereas no degradation was seen when HK was counts for the inhibition of proteolysis in plasma. incubated with recombinant enzymatically inactive PR3 (Fig. 2B, lane 4). Preincubation of PR3 with the anti-PR3 Ab could partially PR3 proteolysis of native HK and of synthetic HK D4 liberates inhibit the PR3-induced proteolysis of HK (Fig. 2C, lanes 2 and 3). a tridecapeptide containing the full bradykinin sequence ␣ 1-AT is the principal inhibitor of PR3 in plasma (12). Incuba- To identify potential kinins liberated from HK by PR3, PR3 was ␣ tion of 1-AT with PR3 before adding HK inhibited PR3-induced incubated with HK, which resulted in the liberation of a peptide The Journal of Immunology 7911

PR3 exerts a hypotensive effect in vivo through both B1 and B2 receptors PR3-kinin and bradykinin were administered i.v. at increasing con-

centrations to wild-type or B2 receptor knockout mice. PR3-kinin induced an instantaneous dose-dependent hypotensive effect equivalent to the effect of bradykinin (Fig. 4A). The effect was

mediated via B2 receptors, as administration of PR3-kinin or bradykinin to B2 receptor knockout mice had no effect on their blood pressure (Fig. 4A) and expression of the B1 receptor is negligible under these experimental conditions. Taken together, the in vitro and in vivo results point to further processing of PR3-kinin into

active derivative(s) capable of B2 receptor stimulation. Transgenic rats TGR-Tie2B1, which overexpress the B1 receptor in the endothelium, and wild-type Hannover Sprague-Dawley rats

were investigated for the in vivo effect of PR3-kinin on B1 receptors. The hypotensive effect was studied after i.v. injection of PR3- kinin and bradykinin at two concentrations. PR3-kinin exerted a

dose-dependent hypotensive effect in TGR-Tie2B1 rats, which was greater than the effect of bradykinin, suggesting that the effect was Downloaded from

partially mediated via binding to the B1 receptor (Fig. 4B). In wild-type rats the effects of PR3-kinin and bradykinin were similar

(Fig. 4B). Pretreatment of the TGR-Tie2B1 rats with HOE-140, a FIGURE 3. PR3-kinin binds to and activates the human B1 receptor selective B2 receptor antagonist, almost totally abrogated the effect but not the B2 receptor. A, PR3-kinin binding to bradykinin receptors. PR3-kinin competition with [3H]des-Arg10-kallidin binding (F) for hu- of bradykinin (Fig. 4C). On the other hand, the hypotensive effect http://www.jimmunol.org/ 3 E of PR3-kinin was only partially inhibited by HOE-140 (Fig. 4C), man kinin B1 receptors and [ H]BK binding ( ) for kinin B2 receptors. PR3-kinin binds to the kinin B1 receptor with relatively high afﬁnity indicating that the effect was partially mediated via B1 receptor whereas it does not interact appreciably with the kinin B2 receptor. B, activity. As a positive control, the selective B1 receptor agonist 10 PR3-kinin-stimulated PI hydrolysis via the B1 receptor. HEK293 cells des-Arg -kallidin was administered. The results indicate that ad- 3 stably expressing B1 receptors were labeled with [ H]myo-inositol and ministration of PR3-kinin i.v. triggers a blood pressure drop me- then assayed for PI hydrolysis in the presence of increasing concentra- diated via both B1 (directly) and B2 receptors (presumably after tions of PR3-kinin. The data are expressed as percentage of maximum, processing). which refers to the stimulation with 1 ␮M des-Arg10-kallidin and are presented as means Ϯ SEM from three separate experiments with each PR3-kinin is processed to bradykinin and des-Arg-bradykinin in data point done in duplicate. by guest on September 23, 2021 plasma

As PR3-kinin bound primarily to B1 receptors in vitro but exerted

an immediate effect via both B1 and B2 receptors in vivo, we in- with a molecular mass of 1492.7 Da (analyzed by MALDI). To vestigated if PR3-kinin was processed in contact with plasma. rule out the possibility of contamination with plasma kallikrein, PR3-kinin incubated with plasma resulted in processing of PR3- PR3 was incubated with a synthetic HK D4 (28-aa-long peptide, kinin to peptides with the molecular mass of bradykinin (1059.4 containing the 9-aa-long bradykinin sequence and its ﬂanking re- Da), des-Arg-bradykinin (903.3 Da), bradykinin-Ser-Ser (1233.7 gions), which showed release of a peptide with the same molecular Da), Met-Lys-bradykinin (1318.4 Da), and Lys-bradykinin-Ser- mass. The amino acid sequence of this peptide was determined by Ser (1361.7 Da), as analyzed by MALDI (data not shown). Pro- cessing commenced at ϳ5 min and after 10 min PR3-kinin was MS/MS and found to be a novel tridecapeptide, NH2-MKRPPG FSPFRSS-COOH, referred to as PR3-kinin. Incubation of HK or completely converted. Plasma alone did not contain any detectable synthetic HK D4 with plasma kallikrein, as a positive control, led bradykinin-like peptides. When PR3-kinin was incubated with to the release of bradykinin (1059.4 Da), as expected, but not PR3- plasma that was preincubated with the three carboxypeptidase in- kinin (data not shown). hibitors separately (EACA, GEMSA, and Plummer’s inhibitor), the processing of PR3-kinin was not inhibited, although no gen-

PR3-kinin binds to and activates the human B1 receptor but eration of des-Arg-bradykinin was detected (data not shown). In does not bind to the B2 receptor contrast, when PR3-kinin was incubated with plasma preincu- PR3-kinin, at concentrations of 10Ϫ10 to 10Ϫ6 M, was added in bated with Complete Mini there was a marked reduction in the a competition binding assay with radiolabeled ligands [3H]des- enzymatic processing of PR3-kinin (data not shown). Taken 10 3 together, these results indicate that the enzyme(s) responsible Arg -kallidin (B1 receptor ligand) or [ H]bradykinin (B2 re- for the proteolytic processing of PR3-kinin in plasma could be ceptor ligand) to HEK293 cells stably expressing the human B1 serine, cysteine, metalloproteases, or calpains, but not car- or B2 receptors (23). PR3-kinin bound with a relatively high Ϯ ϫ boxypeptidase B or N. afﬁnity to the human B1 receptor (IC50 value of 6.3 0.4 10Ϫ9 M, Fig. 3A), whereas it did not interact appreciably with the human B receptor (IC value at Ͼ1 ϫ 10Ϫ6 M, Fig. 3A). PR3-kinin is enzymatically processed to bradykinin and 2 50 des-Arg-bradykinin by plasma kallikrein in vitro Cells stably expressing the B1 receptor were radiolabelled with [3H]myo-inositol and incubated with increasing concentrations of PR3-kinin incubated with plasma kallikrein resulted in rapid PR3-kinin. Data show that PR3-kinin induced PI hydrolysis in a (within 30 s) processing of PR3-kinin to peptides with the molec- dose-dependent manner (Fig. 3B). The results indicate that PR3- ular mass of bradykinin, des-Arg-bradykinin, and Met-Lys-bradykinin binds to and activates the B1 receptor in vitro. kinin, as analyzed by MALDI (data not shown). This breakdown 7912 PROTEINASE 3 RELEASES A NOVEL VASOACTIVE KININ Downloaded from http://www.jimmunol.org/ by guest on September 23, 2021

FIGURE 4. In vivo administration of PR3-kinin stimulates both the B1 and the B2 receptor. A, Wild-type (WT) and B2 receptor knockout (B2KO) mice were injected with cumulative doses of PR3-kinin or bradykinin (0.1, 1, 10, 50, 100 ␮g/kg) and the drop in mean arterial blood pressure (delta MAP, in

mmHg) was measured. The hypotensive effect of PR3-kinin was equivalent to that of bradykinin. The effect was mediated via the B2 receptor as B2KO mice did not respond to PR3-kinin or bradykinin. B, Wild-type rats (WT, Sprague-Dawley) and transgenic rats overexpressing B1 receptors (TGR-Tie2B1) were injected with two different concentrations of PR3-kinin or bradykinin (1 and 10 ␮g/kg) and the drop in mean arterial blood pressure (delta MAP, in ␮ mmHg) was measured. At 10 g/kg PR3-kinin was more efﬁcient than bradykinin in lowering the blood pressure in the TGR-Tie2B1 rats, suggesting that ␮ PR3-kinin induces potent B1 receptor activity. C,AsinB, WT and TGR-Tie2B1 rats were injected with PR3-kinin or bradykinin (10 g/kg). The speciﬁc

effect on B2 receptors was demonstrated by blocking with HOE-140 (HOE) and the drop in mean arterial blood pressure (delta MAP, in mmHg) was measured. PR3-kinin had a more potent hypotensive effect in TGR-Tie2B1 rats than bradykinin and the effect was only partially abolished by 10 ␮ HOE-140, suggesting a B1 receptor-mediated effect. As a positive control, des-Arg -kallidin (0.1 g/kg, a B1 receptor agonist) was administered

to both WT and TGR-Tie2B1 rats.

would not be expected to account for the processing of PR3-kinin lane 4) when compared with HK incubated with the neutrophil detected in normal plasma, as plasma kallikrein is not present in buffer in the absence of neutrophil extracts (Fig. 5, lane 1). normal plasma but circulates in inactive prekallikrein form (1). When proteinase 3 was immunoadsorbed from the neutrophil extracts the proteolysis was markedly reduced (Fig. 5, lanes 3 and PR3 in neutrophil extracts induces HK proteolysis 5). No difference was noted between patients and controls with Neutrophil extracts were obtained from five of the vasculitis pa- regard to the breakdown pattern induced by neutrophil extracts tients (nos. 1, 2, 3, 6, and 7, four with severe ANCA-positive before and after immunoadsorption. Taken together, these results vasculitis) and five controls and examined to investigate if PR3 indicate that PR3-mediated release of PR3-kinin may occur under was involved in HK proteolysis and if differences in this respect physiological conditions, and that this proteolytic process could be could be noted between samples from patients and controls. The enhanced under pathophysiological conditions. neutrophil extracts were incubated with purified HK. Neutrophil extracts from patients treated with Triton X-100 induced break- PR3 in neutrophil extracts induces release of bradykinin-like down of HK into four bands resembling the physiological pattern peptides from HK of proteolysis (Fig. 5, lane 2). Neutrophil extracts from controls Neutrophils extracts from vasculitis patients (nos. 1, 2, 3, 6, and 7) induced a similar breakdown of HK (data not shown). Untreated and controls (n ϭ 5) treated with Triton X-100 induced release of neutrophil extracts (in the absence of Triton X-100) from both bradykinin or bradykinin-like peptide(s) from HK (Fig. 5, lane 2), patients and controls induced minimal breakdown of HK (Fig. 5, whereas untreated neutrophil extracts from the same patients and The Journal of Immunology 7913

FIGURE 5. PR3 in neutrophil extracts induced breakdown of HK. Lane 1, HK (100 ␮g/ ml) with neutrophil buffer without neutrophils. Lane 2, Triton X-100-treated neutrophil extracts from a patient (no. 1) incubated with HK, resulting in breakdown of HK. Lane 3, Triton X-100-treated neutrophil extract from the same patient, in which PR3 had been immunoadsorbed, incubated with HK, resulting in partial breakdown of HK. Lane 4, Untreated neutrophil extracts from the same patient incubated with HK, resulting in almost no breakdown. Lane 5, Untreated neutrophil extract from the same patient, in which PR3 had been immunoadsorbed, incubated with HK, resulting in almost no breakdown of HK. The gel shown is representative for all tested patients (n ϭ 5) and controls (n ϭ 5). No difference in the breakdown pattern of HK was observed between the patients and the controls. The samples were run on the same gel. Results of Downloaded from the ELISA for bradykinin levels (in pg/ml, median (range)) are shown in appropriate lanes under the immunoblot ﬁgure. Neutrophil extracts were incubated with HK (5 ␮g/ml). No differences in the levels of bradykinin between the patients and the controls were observed, al- http://www.jimmunol.org/ though a tendency to higher bradykinin levels was noted when incubating the patients’ Triton X-100-treated neutrophil extracts with HK.

controls did not ( p ϭ 0.002; Fig. 5, lane 4). Bradykinin levels did flammatory conditions, in which activated neutrophils infiltrate ex- not differ significantly between HK-incubated neutrophil extracts travascular spaces, neutrophil-derived PR3 in membrane-bound or from patients and controls. When proteinase 3 was immunoad- secreted form could generate PR3-kinin in a kallikrein-indepen-

␣ by guest on September 23, 2021 sorbed from the neutrophil extract of patients and controls the dent manner and that in tissues where 1-AT levels are presumably ␣ release of bradykinin or bradykinin-like peptide(s) was signiﬁ- low, and/or 1-AT activity is compromised due to oxidation of a cantly reduced (Fig. 5, lane 3). crucial methionine residue (33), liberation of kinins could proceed PR3 levels were measured in Triton X-100-treated neutrophil in a relatively uninhibited manner. extracts from the ﬁve vasculitis patients and the controls. No dif- The results suggest that PR3-kinin can exert an effect both via ferences were found in the levels between the patients (median, 3.9 the B1 receptor (directly) and the B2 receptor (after processing to ␮g/ml; range 2.7–7.2 ␮g/ml) and the controls (median, 4.8 ␮g/ml; bradykinin). In plasma, PR3-kinin is processed to bradykinin and range, 1.6–7.3 ␮g/ml; p ϭ 0.69). The results indicate that PR3 in des-Arg-bradykinin within 5–10 min. Although our analysis has neutrophil extracts leads to the breakdown of HK and subsequent not unequivocally assigned the lacking Arg residue to the C ter- 9 release of bradykinin or bradykinin-like peptide(s), and no differ- minus of bradykinin, des-Arg -bradykinin, which binds to B1 re- ences in this respect were noted in neutrophils from patients and ceptors, may have been generated, whereas des-Arg1-bradykinin controls. does not bind to B1 and B2 receptors (2). The effect of PR3-kinin

in vivo on both B1 and B2 receptors was immediate, suggesting

Discussion that conversion of PR3-kinin to the form active on B2 receptors

This study presents a novel mechanism of kinin pathway activation occurs very rapidly in vivo. Murine B2 receptors differ pharmaco- in which HK breakdown by PR3 was demonstrated in the ﬂuid logically from human B2 receptors in their response to antagonists phase, in the absence of endothelial cells or neutrophils. More (2, 34). Thus, species-speciﬁc properties may also contribute to the importantly, release of a novel kinin, termed PR3-kinin, occurred effect of PR3-kinin in mice. in the absence of plasma kallikrein, the known physiological en- Plasma kallikrein was also capable of proteolytic cleavage of zymatic activator of this system. PR3-kinin bound to and activated PR3-kinin to bradykinin in vitro. In normal plasma active kal- the human B1 receptor but did not bind to the human B2 receptor likrein would not be present (1). However, on activated neutrophils in vitro. PR3-kinin was as potent as the B2 agonist bradykinin in and endothelial cells, where plasma kallikrein is active (35), lowering blood pressure in wild-type mice, an effect that was ab- such proteolytic breakdown of PR3-kinin could potentially occur sent in B2 knockout mice. PR3-kinin was more potent than bra- in vivo. dykinin in B1 receptor overexpressing TGR-Tie2B1 rats, suggest- Previous studies have shown that stimulated neutrophils induced ing that the in vivo hypotensive effects were exerted via both B1 the generation of biologically active kinins from HK (36–38), and B2 receptors. These results indicate that PR3-kinin is a B1 whereas nonstimulated neutrophils did not hydrolyze HK (39). receptor agonist and that it is converted in vivo to a B2 receptor Neutrophil-derived tissue kallikrein participated in kinin release as agonist. This novel kinin-generating pathway is controlled by the reaction was blocked by antihuman urinary kallikrein but not ␣ 1-AT and would thus circumvent the established inhibitory mech- by antineutrophil elastase Ab (36). The generated kinins were an- anisms dominated by C1 inhibitor. We envisage that during in- alyzed by HPLC and radioimmunoassay and shown to have similar 7914 PROTEINASE 3 RELEASES A NOVEL VASOACTIVE KININ

FIGURE 6. HK sequence cleaved by PR3. The PR3-kinin sequence is marked in gray. h, Hydrophobic; neg, negative; pos, positive; h-phil, hydro- philic; n, neutral.

retention times to bradykinin and Met-Lys-bradykinin. The bio- We have shown a novel function for PR3 in inflammation, logical activity was mediated by kinin B2 receptors (36). We have namely generation of the vasoactive kinin PR3-kinin. The typical taken these results further and demonstrated, using Triton X-100 pathological finding in vasculitis consists of neutrophil infiltrates lyzed neutrophils, that PR3, in a dose-dependent, specific, and in and around vessels and in tissues. Activated neutrophils infil- physiological manner, is almost exclusively responsible for HK trating affected tissues secrete granular components including PR3. proteolysis and that the product of this reaction is a novel kinin This could create a microenvironment rich in cytokine-activated peptide that we have termed PR3-kinin. This tridecapeptide cor- neutrophils that express PR3 on their membrane and secrete PR3 Downloaded from responds to Met-Lys-bradykinin with two additional serine resi- into the blood vessels and tissues. PR3-induced activation of the dues at the C terminus. By studying the reaction of purified PR3 (in kinin pathway could occur at these sites and possibly in secluded ␣ native and recombinant form) with HK and synthetic HK D4, we environments that do not allow 1-AT entry. Thus PR3-kinin is could rule out the participation of contaminating enzymes, such as formed and binds to the B1 receptor followed by capillary leakage elastase or plasma kallikrein, in these proteolytic processes. Fur- (2). This will allow exposure to plasma enzymes that rapidly con- thermore, the Abs used to immunoadsorb PR3 from neutrophil vert PR3-kinin into the active B2 agonist bradykinin, further en- http://www.jimmunol.org/ extracts, and to block PR3 activity in specificity assays, have been hancing the inflammatory response. shown to be specific for PR3 and do not cross-react with other proteins in neutrophil azurophilic granules (29) or with kallikreins, Acknowledgments as demonstrated in this study. The authors thank the Swegene Centre for Integrative Biology at Lund In addition to PR3, other neutrophil proteases, such as elastase University and especially Peter James, Liselotte Andersson, and Mats and cathepsins, have been shown to cleave HK and liberate pep- Mågård for their excellent assistance with mass spectrometry. We also tides with bradykinin-like properties (40–43). Neutrophil elastase thank Caroline Sandeń for her excellent assistance with the bradykinin releases a peptide from HK termed E-kinin SLMKRPPGFSP receptor phosphoinositide hydrolysis assay as well as Lena Gunnarson for FRSSRI (the bradykinin sequence is underlined). E-kinin and her excellent technical assistance. by guest on September 23, 2021 two synthetic E-kinin-like peptides (SLMKRPPGFSPFRSS and SLMKRPPGFSPFR) are longer than bradykinin and PR3-kinin Disclosures and are capable of lowering blood pressure in an animal model (41). The authors have no financial conflicts of interest. The preferential amino acid sequence for proteinase 3 cleavage consists of hydrophobic amino acids at the P1, P3, and P4 posi- Ј References tions, negatively charged amino acids at P2 and P2 , and positively 1. Colman, R. W., and A. H. Schmaier. 1997. Contact system: a vascular biology charged amino acids at P1Ј and P3Ј. Although this represents the modulator with anticoagulant, profibrinolytic, antiadhesive, and proinflammatory ideal sequence, other amino acids may be found at the cleavage attributes. Blood 90: 3819–3843. 2. Leeb-Lundberg, L. M., F. Marceau, W. Muller-Esterl, D. J. Pettibone, and sites (44, 45). The cleavage sites on HK do not fully match the B. L. Zuraw. 2005. International union of pharmacology, XLV: Classification of consensus sequence established for PR3 (46) (Fig. 6); however, the kinin receptor family: from molecular mechanisms to pathophysiological con- our results clearly show that PR3 releases PR3-kinin from HK, sequences. Pharmacol. Rev. 57: 27–77. 3. Phagoo, S. B., S. Poole, and L. M. Leeb-Lundberg. 1999. Autoregulation of allowing cleavage at nonconsensus sites. bradykinin receptors: agonists in the presence of interleukin-1␤ shift the reper- Patients with vasculitides exhibit activation of the kinin system toire of receptor subtypes from B2 to B1 in human lung fibroblasts. Mol. Phar- as demonstrated by degradation of HK and elevated plasma levels macol. 56: 325–333. 4. Kahn, R., H. Herwald, W. Muller-Esterl, R. Schmitt, A. C. Sjogren, L. Truedsson, of bradykinin (4). In a previous study we showed that patients with and D. Karpman. 2002. Contact-system activation in children with vasculitis. both mild (Henoch Scho¨nlein purpura) and severe (Wegener’s Lancet 360: 535–541. granulomatosis) vasculitis had kinin system activation whereas 5. Stadnicki, A., M. Gonciarz, T. J. Niewiarowski, J. Hartleb, M. Rudnicki, N. B. Merrell, R. A. Dela Cadena, and R. W. Colman. 1997. Activation of plasma controls did not (4). In the present study not all patients had contact and coagulation systems and neutrophils in the active phase of ulcerative ANCA-related vasculitis. Patients with ANCA-associated vascu- colitis. Dig. Dis. Sci. 42: 2356–2366. litis had higher levels of PR3 in plasma than did controls (16, 17) 6. van der Geld, Y. M., P. C. Limburg, and C. G. Kallenberg. 2001. Proteinase 3, Wegener’s autoantigen: from gene to antigen. J. Leukocyte Biol. 69: 177–190. and the expression of PR3 on the membrane of leukocytes corre- 7. Csernok, E., M. Ernst, W. Schmitt, D. F. Bainton, and W. L. Gross. 1994. Ac- lates with disease activity in patients with ANCA-associated vas- tivated neutrophils express proteinase 3 on their plasma membrane in vitro and in culitis (Wegener’s granulomatosis) (8). However, as presented in vivo. Clin. Exp. Immunol. 95: 244–250. 8. Muller Kobold, A. C., C. G. Kallenberg, and J. W. Tervaert. 1998. Leucocyte the present study, even neutrophil extracts from healthy controls membrane expression of proteinase 3 correlates with disease activity in patients contained PR3 capable of cleaving HK. We did not detect differ- with Wegener’s granulomatosis. Br. J. Rheum. 37: 901–907. 9. Leid, R. W., B. E. Ballieux, I. van der Heijden, C. Kleyburg-van der Keur, ences in PR3 levels or activity in neutrophil extracts from patients E. C. Hagen, L. A. van Es, F. J. van der Woude, and M. R. Daha. 1993. Cleavage or controls. Thus, PR3-induced activation of the kinin pathway and inactivation of human C1 inhibitor by the human leukocyte proteinase, pro- may also occur at low PR3 levels under physiological conditions; teinase 3. Eur. J. Immunol. 23: 2939–2944. 10. Taekema-Roelvink, M. E., C. Van Kooten, E. Heemskerk, W. Schroeijers, and however, it may be abundant at sites of massive inflammation and M. R. Daha. 2000. Proteinase 3 interacts with a 111-kD membrane molecule of neutrophil accumulation around vessels. human umbilical vein endothelial cells. J. Am. Soc. Nephrol. 11: 640–648. The Journal of Immunology 7915

11. Yang, J. J., R. Kettritz, R. J. Falk, J. C. Jennette, and M. L. Gaido. 1996. Apo- 29. Sommarin, Y., N. Rasmussen, and J. Wieslander. 1995. Characterization of ptosis of endothelial cells induced by the neutrophil serine proteases proteinase 3 monoclonal antibodies to proteinase-3 and application in the study of epitopes for and elastase. Am. J. Pathol. 149: 1617–1626. classical anti-neutrophil cytoplasm antibodies. Exp. Nephrol. 3: 249–256. 12. Rao, N. V., N. G. Wehner, B. C. Marshall, W. R. Gray, B. H. Gray, and 30. Selga, D., M. Segelmark, J. Wieslander, L. Gunnarsson, and T. Hellmark. 2004. J. R. Hoidal. 1991. Characterization of proteinase-3 (PR-3), a neutrophil serine Epitope mapping of anti-PR3 antibodies using chimeric human/mouse PR3 re- proteinase: structural and functional properties. J. Biol. Chem. 266: 9540–9548. combinant proteins. Clin. Exp. Immunol. 135: 164–172. 13. Kamesh, L., L. Harper, and C. O. Savage. 2002. ANCA-positive vasculitis. J. Am. 31. Ohlsson, S., T. Hellmark, K. Pieters, G. Sturfelt, J. Wieslander, and Soc. Nephrol. 13: 1953–1960. M. Segelmark. 2005. Increased monocyte transcription of the proteinase 3 gene 14. Kallenberg, C. G., P. Heeringa, and C. A. Stegeman. 2006. Mechanisms of dis- in small vessel vasculitis. Clin. Exp. Immunol. 141: 174–182. ease: pathogenesis and treatment of ANCA-associated vasculitides. Nat. Clin. 32. Bhoola, K. D., C. D. Figueroa, and K. Worthy. 1992. Bioregulation of kinins: Pract. Rheumatol. 2: 661–670. kallikreins, kininogens, and kininases. Pharmacol. Rev. 44: 1–80. 15. Savage, C. O., L. Harper, and D. Adu. 1997. Primary systemic vasculitis. Lancet 349: 553–558. 33. Swaim, M. W., and S. V. Pizzo. 1988. Methionine sulfoxide and the oxidative 16. Henshaw, T. J., C. C. Malone, J. E. Gabay, and R. C. Williams, Jr. 1994. Ele- regulation of plasma proteinase inhibitors. J. Leukocyte Biol. 43: 365–379. vations of neutrophil proteinase 3 in serum of patients with Wegener’s granulo- 34. Hess, J. F., J. A. Borkowski, T. Macneil, G. Y. Stonesifer, J. Fraher, matosis and polyarteritis nodosa. Arthritis Rheum. 37: 104–112. C. D. Strader, and R. W. Ransom. 1994. Differential pharmacology of cloned 17. Abdgawad, M., T. Hellmark, L. Gunnarsson, K. W. Westman, and M. Segelmark. human and mouse B2 bradykinin receptors. Mol. Pharmacol. 45: 1–8. 2006. Increased neutrophil membrane expression and plasma level of proteinase 35. Shariat-Madar, Z., F. Mahdi, and A. H. Schmaier. 2002. Assembly and activation 3 in systemic vasculitis are not a consequence of the Ϫ564 A/G promotor poly- of the plasma kallikrein/kinin system: a new interpretation. Int. Immunopharma- morphism. Clin. Exp. Immunol. 145: 63–70. col. 2: 1841–1849. 18. Skold, S., B. Rosberg, U. Gullberg, and T. Olofsson. 1999. A secreted proform 36. Stuardo, M., C. B. Gonzalez, F. Nualart, M. Boric, J. Corthorn, K. D. Bhoola, and of neutrophil proteinase 3 regulates the proliferation of granulopoietic progenitor C. D. Figueroa. 2004. Stimulated human neutrophils form biologically active cells. Blood 93: 849–856. kinin peptides from high and low molecular weight kininogens. J. Leukocyte Biol. 19. Skold, S., L. Zeberg, U. Gullberg, and T. Olofsson. 2002. Functional dissociation 75: 631–640. between proforms and mature forms of proteinase 3, azurocidin, and granzyme B

37. Movat, H. Z., F. M. Habal, and D. R. Macmorine. 1976. Generation of a vaso- Downloaded from in regulation of granulopoiesis. Exp. Hematol. 30: 689–696. active peptide by a neutral protease of human neutrophil leukocytes. Agents Ac- 20. Fernlund, P., G. Fex, A. Hanson, J. Stenflo, and B. Lundh. 1991. Laurells Klinisk tions 6: 183–190. Kemi. In Plasmaproteiner, 6th Ed. Studentlitteratur, Lund, Sweden, pp. 387–389. 21. Muller-Esterl, W., D. A. Johnson, G. Salvesen, and A. J. Barrett. 1988. Human 38. Wintroub, B. U., E. J. Goetzl, and K. F. Austen. 1974. A neutrophil-dependent pathway for the generation of a neutral peptide mediator: partial characterization kininogens. Methods Enzymol. 163: 240–256. ␣ 22. Kaufmann, J., M. Haasemann, S. Modrow, and W. Muller-Esterl. 1993. Struc- of components and control by 1-antitrypsin. J. Exp. Med. 140: 812–824. tural dissection of the multidomain kininogens: fine mapping of the target 39. Figueroa, C. D., M. Stuardo, C. B. Gonzalez, F. Nualart, J. Corthorn, and epitopes of antibodies interfering with their functional properties. J. Biol. Chem. K. D. Bhoola. 1999. Hydrolysis of kininogens by degranulated human neutrophils and analysis of bradykinin as chemotactic factor for cells isolated from peripheral 268: 9079–9091. http://www.jimmunol.org/ 23. Enquist, J., C. Skroder, J. L. Whistler, and L. M. Leeb-Lundberg. 2007. Kinins blood. Immunopharmacology 43: 211–217. promote B2 receptor endocytosis and delay constitutive B1 receptor endocytosis. 40. Imamura, T., J. Potempa, and J. Travis. 2004. Activation of the kallikrein-kinin Mol. Pharmacol. 71: 494–507. system and release of new kinins through alternative cleavage of kininogens by 24. Borkowski, J. A., R. W. Ransom, G. R. Seabrook, M. Trumbauer, H. Chen, microbial and human cell proteinases. Biol. Chem. 385: 989–996. R. G. Hill, C. D. Strader, and J. F. Hess. 1995. Targeted disruption of a B2 41. Imamura, T., S. Tanase, I. Hayashi, J. Potempa, A. Kozik, and J. Travis. 2002. bradykinin receptor gene in mice eliminates bradykinin action in smooth muscle Release of a new vascular permeability enhancing peptide from kininogens by and neurons. J. Biol. Chem. 270: 13706–13710. human neutrophil elastase. Biochem. Biophys. Res. Commun. 294: 423–428. 25. Cayla, C., M. Todiras, R. Iliescu, V. V. Saul, V. Gross, B. Pilz, G. Chai, 42. Puzer, L., J. Vercesi, M. F. Alves, N. M. Barros, M. S. Araujo, M. Aparecida Juliano, V. F. Merino, J. B. Pesquero, O. C. Baltatu, and M. Bader. 2007. Mice deficient M. L. Reis, L. Juliano, and A. K. Carmona. 2005. A possible alternative mechanism for both kinin receptors are normotensive and protected from endotoxin-induced of kinin generation in vivo by cathepsin L. Biol. Chem. 386: 699–704. hypotension. FASEB J. 21: 1689–1698. 43. Kleniewski, J., and V. Donaldson. 1988. Granulocyte elastase cleaves human 26. Merino, V. F., M. Todiras, L. A. Campos, V. Saul, E. Popova, O. C. Baltatu, by guest on September 23, 2021 high molecular weight kininogen and destroys its clot-promoting activity. J. Exp. J. B. Pesquero, and M. Bader. 2008. Increased susceptibility to endotoxic shock Med. 167: 1895–1907. in transgenic rats with endothelial overexpression of kinin B1 receptors. J. Mol. Med. 86: 791–798. 44. Korkmaz, B., T. Moreau, and F. Gauthier. 2008. Neutrophil elastase, proteinase 27. Jennette, J. C., R. J. Falk, K. Andrassy, P. A. Bacon, J. Churg, W. L. Gross, 3 and cathepsin G: physicochemical properties, activity and physiopathological E. C. Hagen, G. S. Hoffman, G. G. Hunder, C. G. Kallenberg, et al. 1994. No- functions. Biochimie 90: 227–242. menclature of systemic vasculitides: proposal of an international consensus con- 45. Hajjar, E., B. Korkmaz, F. Gauthier, B. O. Brandsdal, V. Witko-Sarsat, and ference. Arthritis Rheum. 37: 187–192. N. Reuter. 2006. Inspection of the binding sites of proteinase 3 for the design of 28. Tapper, H., A. Karlsson, M. Morgelin, H. Flodgaard, and H. Herwald. 2002. a highly specific substrate. J. Med. Chem. 49: 1248–1260. Secretion of heparin-binding protein from human neutrophils is determined by its 46. Stryer, L. 1995. Biochemistry. In Molecular Design of Life, 4th Ed. Freeman, localization in azurophilic granules and secretory vesicles. Blood 99: 1785–1793. New York, pp. 17–45.