Tissue Inhibitor of Metalloproteinase-1 Is Cleaved and Inactivated Extracellularly by α-Chymase

This information is current as Brendon T. Frank, J. Caleb Rossall, George H. Caughey and of September 24, 2021. Kenneth C. Fang J Immunol 2001; 166:2783-2792; ; doi: 10.4049/jimmunol.166.4.2783 http://www.jimmunol.org/content/166/4/2783 Downloaded from

References This article cites 56 articles, 23 of which you can access for free at: http://www.jimmunol.org/content/166/4/2783.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 24, 2021 *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 © 2001 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Mast Cell Tissue Inhibitor of Metalloproteinase-1 Is Cleaved and Inactivated Extracellularly by ␣-Chymase1

Brendon T. Frank,* J. Caleb Rossall,* George H. Caughey,*† and Kenneth C. Fang2*†

We previously reported that mast cell ␣-chymase cleaves and activates progelatinase B (progel B). Outside of cells, progel B is complexed with tissue inhibitor of metalloproteinase (TIMP)-1, which hinders zymogen activation and inhibits activity of mature forms. The current work demonstrates that dog BR mastocytoma cells, HMC-1 cells, and murine bone marrow-derived mast cells secrete TIMP-1 whose electrophoretic profile in supernatants suggests degranulation-dependent proteolysis. ␣-Chymase cleaves uncomplexed TIMP-1, reducing its ability to inhibit gel B, whereas has no effect. Sequencing of TIMP-1’s ␣-chymase- mediated cleavage products reveals hydrolysis at Phe12-Cys13 and Phe23-Val24 in loop 1 and Phe101-Val102 and Trp105-Asn106 in loop 3 of the NH2-terminal domain. TIMP-1 in a ternary complex with progel B and neutrophil gelatinase-associated lipocalin is ␣ ␣ also susceptible to -chymase cleavage, yielding products like those resulting from processing of free TIMP-1. Thus, -chymase Downloaded from cleaves free and gel B-bound TIMP-1. Incubation of the progel B-TIMP-1-neutrophil gelatinase-associated lipocalin complex with ␣-chymase increases gel B activity 2- to 5-fold, suggesting that ␣-chymase activates progel B whether it exists as free monomer or as a complex with TIMP-1. Furthermore, inhibition of ␣-chymase blocks degranulation-induced TIMP-1 processing (absent in ␣-chymase-deficient HMC-1 cells). Purified ␣-chymase processes TIMP-1 in BR supernatants, generating products like those induced by degranulation. In summary, these results suggest that controlled exocytosis of mast cell ␣-chymase activates progel B

even in the presence of TIMP-1. This is the first identification of a protease that overcomes inhibition by bound TIMP-1 to activate http://www.jimmunol.org/ progel B without involvement of other proteases. The Journal of Immunology, 2001, 166: 2783–2792.

ctivated mast cells respond to allergen, tumor invasion, inhibitors of metalloproteinases (TIMPs). Proteolytic activation of and fibrogenic injury by releasing tryptase and ␣-chy- pro-MMPs involves one or more activator proteases which remove A mase, which are granule-associated tryptic and chymo- the propeptide domain containing a critical Cys residue, thus dis- tryptic serine proteases, respectively. Upon degranulation, both rupting the cysteine switch which confers latency (9). proteases participate in proteolytic and nonproteolytic pathways TIMPs inhibit activity of active MMP species and block activation regulating matrix protein degradation, receptor activation, peptide of pro-MMPs by forming TIMP-MMP complexes in a 1:1 molar inactivation, and fibroblast, smooth muscle, or epithelial cell mi- by guest on September 24, 2021 ratio. NH2-terminal TIMP domains occupy the MMP active site togenesis (1). Tryptase and ␣-chymase may also act in a broader and COOH-terminal domains confer binding specificity through range of homeostatic and pathologic tissue remodeling processes interactions with COOH-terminal MMP domains (10). 2ϩ 2ϩ via interactions with a family of Ca - and Zn -dependent matrix Mast cells may regulate local MMP activity by contributing zy- 3 metalloproteinases (MMPs) (2–7). Secreted or cell surface-asso- mogen-activating serine proteases and by secreting pro-MMPs. ciated membrane-type MMPs participate in physiologic pathways Our previous work demonstrates that mast cells secrete progelati- such as embryonic development, organ morphogenesis, angiogen- nase B (progel B; MMP-9), which is activated extracellularly by esis, and wound healing, and also contribute to the pathogenesis of degranulated ␣-chymase upon cleavage of the catalytic domain at arthritis, cancer, cardiovascular disease, and lung fibrosis (8). Post- two sites (2–4). ␣-Chymase-dependent activation of progel B in translational regulation of MMP activity depends upon conversion vivo is hypothesized to switch nonangiogeneic tissues to an an- of proenzymes to mature active forms and their inhibition by tissue giogenic phenotype in premalignant lesions in a murine model of epithelial carcinogenesis (11). In addition to progel B, ␣-chymase also directly cleaves and activates procollagenase (MMP-1) and *Cardiovascular Research Institute and †Department of Medicine, University of Cal- ifornia, San Francisco, CA 94143 prostromelysin (MMP-3) (6, 7). Mast cells also express these two Received for publication October 4, 2000. Accepted for publication December MMPs plus progelatinase A (MMP-2), which shares in vitro sub- 7, 2000. strate specificity with gelatinase B (gel B) (4, 12, 13). The costs of publication of this article were defrayed in part by the payment of page Since cells secrete pro-MMPs bound to TIMPs in a complex, charges. This article must therefore be hereby marked advertisement in accordance zymogen activation likely requires prior TIMP processing which with 18 U.S.C. Section 1734 solely to indicate this fact. precedes proteolytic activation of the MMP moiety. TIMP-1 binds 1 This work was supported by Grants HL-03345 and HL-24136 from the National Institutes of Health. K.C.F. is the recipient of a Mentored Clinical Scientist Devel- to progel B, which is secreted as a monomer or dimer, or in di- opment Award from the National Institutes of Health. sulfide-mediated linkage with neutrophil gelatinase-associated li- 2 Address correspondence and reprint requests to Dr. Kenneth C. Fang, Box 0911, pocalin (NGAL) to form a ternary complex of progel B monomer, Cardiovascular Research Institute, University of California, San Francisco, CA TIMP-1 and NGAL (progel B-TIMP-1-NGAL) (14, 15). Unbound 94143-0911. E-mail address: [email protected] TIMP-1 may be inactivated in vitro by cleavage, degradation, or 3 Abbreviations used in this paper: MMP, matrix metalloproteinase; AAPF-CMK, Ala-Ala-Pro-Phe chloromethyl ketone; DNP-PGCHAK, DNP-Pro-Gly-Cys(Me)-His- chemical modification via proteolytic and nonproteolytic mecha- Ala-Lys(N-Me-Abz)-NH2; gel B, gelatinase B; HNE, human neutrophil ; KL, nisms involving serine or thiol proteases and reactive oxygen spe- recombinant canine kit ligand; MBMMC, murine bone marrow-derived mast cells; NGAL, neutrophil gelatinase-associated lipocalin; progel B, progelatinase B; TIMP, cies, respectively (16–19). However, mechanisms which regulate tissue inhibitor of metalloproteinase; rh, recombinant human. processing of MMP-bound TIMP-1 remain unclear. In addition to

Copyright © 2001 by The American Association of Immunologists 0022-1767/01/$02.00 2784 ␣-CHYMASE CLEAVAGE OF TIMP-1 its ability to block pro-MMP activation and inhibit activity of ma- fore further analysis. To determine the electrophoretic profile of reaction ture MMPs, TIMP-1 also demonstrates erythroid-potentiating ac- products, aliquots were subjected to SDS-PAGE using 16% or 4–20% tivity, stimulates steroidogenesis, regulates mitogenesis, and con- gradient Tris-glycine gels (Invitrogen) under reducing conditions, with de- tection of proteins by Coomassie Blue R250 (Fisher Scientific, Tustin, CA) trols apoptosis via mechanisms which are independent of its MMP or immunoblotting using polyclonal rabbit anti-TIMP-1 or polyclonal rab- inhibitory activity (10). Thus, processing of TIMP-1 in the inflam- bit anti-MMP-9 Abs (Triple Point Biologics, Portland, OR). matory milieu may not only alter the protease:antiprotease balance TIMP-1 activity to favor proteolysis, but also attenuate or abolish its non-MMP inhibitory effects on specific cell populations. We report here that MMP inhibitory activity of rhTIMP-1 was determined by incubating intact ␣ mast cell ␣-chymase cleaves free TIMP-1 and processes the progel or -chymase-processed inhibitor with active gel B monomer (Calbio- chem). rhTIMP-1 was incubated alone or in the presence of various con- B-TIMP-1 complex by inactivating bound TIMP-1 and activating centrations of ␣-chymase in reaction buffer at 37°C for different time pe- progel B, without involvement of other secreted proteases, and that riods. Reactions were stopped by incubation with Ala-Ala-Pro-Phe- degranulation of ␣-chymase regulates processing of TIMP-1 se- chloromethyl ketone (AAPF-CMK; Systems Products, Livermore, creted by mast cells. CA) at 22°C for 10 min at a final concentration of 25 ␮M. Active gel B monomer was incubated alone or with aliquots of each reaction mix in a 1:1 molar ratio at 22°C for 10 min. The gel B activity in solution was Materials and Methods determined in a final volume of 200 ␮l of reaction buffer containing 25 ␮M Cell culture AAPF-CMK in 96-well plates (Becton Dickinson, Franklin Lakes, NJ) with 10 ␮M DNP-Pro-Cha-Gly-Cys(Me)-His-Ala-Lys(N-Me-Abz)-NH BR dog mastocytoma cells or HMC-1 cells (obtained from J. Butterfield, 2 (Cha ϭ b-cyclohexylalanyl; Abz ϭ 2-aminobenzoyl(anthraniloyl)] (DNP- Mayo Clinic, Rochester, MN) were maintained in continuous suspension PGCHAK; Calbiochem) (23). Fluorescent products released by cleavage of culture in DME-H16 medium supplemented with 2% bovine calf serum (2) Downloaded from substrate were detected at different times at excitation and emission wave- or IMDM with 2% FCS (20), respectively, as previously described. Murine lengths of 360 nm and 460 nm, respectively, using a CytoFluor 2350 fluo- bone marrow-derived mast cells (MBMMC) obtained from C57BL/6 mice rometer (Millipore, Danbury, CT) and a sensitivity parameter of 6. (The Jackson Laboratory, Bar Harbor, ME) were cultured in RPMI 1640 medium supplemented with 10% bovine calf serum and 50% medium con- NH2-terminal sequencing ditioned by WEHI-3B cells, as previously described (21). Cells were main- tained at a final concentration of 1 ϫ 106 cells/ml and incubated at 37°C in Amino acid sequence was determined by the Biomolecular Resource Cen- humidified 5% CO2 and 95% air. Cells were harvested by centrifuging at ter at the University of California San Francisco. The reaction mixture 500 ϫ g for 5 min, washing three times in Ca2ϩ-and Mg2ϩ-free PBS, and containing the TIMP-1 products of ␣-chymase cleavage was electropho- http://www.jimmunol.org/ resuspending in serum-free culture medium to a final concentration of 10– resed onto a 4–20% Tris-glycine polyacrylamide gel and blotted onto se- 15 ϫ 106 cells/ml. quencing grade polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA) in 10 mM 3-[cyclohexylamino]-1-propanesulfonic acid buffer con- RNA blotting taining 10% methanol. Protein bands identified by Coomassie blue staining ϩ of the membrane were excised and subjected to Edman degradation using Poly(A) RNA was isolated from BR mastocytoma cells incubated alone a Procise 491 Protein Sequencer (PE Biosystems, Foster City, CA). Protein or in the presence of 100 ng/ml recombinant canine KL (kindly provided ϩ sequence alignments were performed using MacVector software (Oxford by Keith Langley, Amgen, Thousand Oaks, CA) using the Poly(A) RNA Molecular Group, Hunt Valley, MD). MicroFast Track extraction kit (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol. RNA blotting and hybridization of random Gelatin zymography primed [␣-32P]ATP-labeled 0.79-kb dog TIMP-1 cDNA (gift from Philip by guest on September 24, 2021 Davies, Merck Research Laboratories, Rahway, NJ) or 2.3-kb dog gel B (3) The gel B activity resulting from ␣-chymase cleavage of the progel cDNA to nylon membranes, posthybridization washings, and autoradiog- B-TIMP-1-NGAL complex was detected by gelatin zymography as previ- raphy were performed as previously described (4). Densitometric data were ously described (2). Briefly, reaction products were electrophoresed obtained by analysis of autoradiographic signals generated by hybridizing through 10% polyacrylamide gels copolymerized with gelatin (1 mg/ml). the blot with labeled probe. To account for possible variations in signal After electrophoresis, gels were washed twice for 30 min in 2.5% Triton intensity due to differing concentrations of mRNA present in each lane, the X-100 and incubated at 37°C for 18 h in 40 mM Tris buffer (pH 7.5) blot was also hybridized with a labeled probe for ␥-actin. Densitometric containing 200 mM NaCl and 10 mM CaCl2. Gels were stained with Coo- data were then compared with control values obtained with the ␥-actin massie blue for 10 min and destained in 10% acetic acid containing 50% probe. methanol. Clear zones of lysis against a blue background indicated gela- tinase activity. Immunoblotting Enzyme assays Aliquots of medium conditioned by mast cells were concentrated by Mi- crocon-10 filters (Amicon, Beverly, MA), subjected to electrophoresis, and Soluble MMP activity was determined by detection of products released blotted onto polyvinylidene difluoride membrane (Polyscreen; NEN Life upon cleavage of fluorescently labeled proteins or synthetic peptides by Science Products, Boston, MA). Membranes were washed in 10 mM Tris- active proteases. The gel B monomer or the progel B-TIMP-1-NGAL com- HCl (pH 7.5) containing 150 mM NaCl and 0.3% Tween 20 and incubated plex was incubated alone or with activator proteases at 37°C for different with polyclonal rabbit anti-TIMP-1 (Chemicon, Temecula, CA; rabbit time periods in 200 ␮l of reaction buffer in 96-well plates with various polyclonal Ab is specific for TIMP-1 and does not cross-react with other concentrations of DNP-PGCHAK (23) or DQ gelatin (fluorescein conju- TIMPs in the manufacturer’s quality control assays) at 22°C for 1 h. After gate from pig skin) (Molecular Probes, Eugene, OR) to measure gel B incubation of membranes with HRP-linked anti-rabbit Ig for 1 h and lu- activity or FITC-conjugated casein (Sigma) to assay stromelysin activity. minol and peroxidase reagents (Phototope-HRP Western Blot Detection Specific activity of active gel B monomer or active stromelysin catalytic kit; New England Biolabs, Beverly, MA) according to the manufacturer’s domain using DQ gelatin or FITC casein was 186 fluorescent units/␮g/min protocols, signals for immunoreactive proteins were visualized on Hyper- or 280 fluorescent units/␮g/min, respectively. Background levels of sub- film ECL (Amersham Pharmacia, Piscataway, NJ). strate cleavage by ␣-chymase, HNE, or were subtracted from in- dividual determinations of soluble activity. Substrate cleavage Degranulation studies Purified recombinant human (rh) TIMP-1 (Chemicon) or the progel B-TIMP-1-NGAL complex (Calbiochem, La Jolla, CA) was incubated ei- Harvested cells were resuspended in serum-free culture medium to a final ther alone or in combination with different concentrations of activator pro- concentration of 10–15 ϫ 106 cells/ml. Cells were incubated at 37°C for teases including: purified rh␣-chymase (22), purified rh␤-tryptase (Pro- different time periods either alone, with 2 ␮M calcium ionophore A23187 mega, Madison, WI), human (HNE; Sigma, St. Louis, (Sigma), or with ionophore in combination with individual inhibitors (Sig- MO), L-1-tosylamido-2-phenylethyl chloromethyl ketone-treated trypsin ma) to identify protease classes involved in the processing of endogenous (Sigma), and purified rh stromelysin (MMP-3) catalytic domain (22 kDa; TIMP-1. Various concentrations of PMSF, aprotinin, or AAPF-CMK were Calbiochem). Reactions were performed in 50 mM Tris-HCl buffer (pH used to block activity, while E64 or phosphoramidon and

7.5) containing 150 mM NaCl, 10 mM CaCl2, and 0.02% NaN3 (reaction BB94 (gift from Marc Navre) were used to inhibit activity of cysteine buffer), incubated at 37°C for various time periods, and stored at 0°C be- proteases or metalloproteases (Affymay Research Institute, Santa Clara, The Journal of Immunology 2785

FIGURE 1. Mast cell expression of TIMP-1. A, Poly(A)ϩ RNA isolated from BR cells incubated in medium alone (C) or in the presence of 100 ng/ml KL was separated on a 1% agarose gel containing 6.1% formaldehyde and transferred to nylon membrane which was hybridized with an [␣-32P]-labeled probe for dog TIMP-1 or gel B. B, Densitometric data were obtained in duplicate experiments by analysis of autoradiographic signals after 96 h and normalized to values obtained with the ␥-actin probe. Values are expressed as a percentage of the control signal obtained from unstimulated cells. C, BR cells were incubated in serum-free medium alone (C) or in the presence of 2 mM or 10 mM PMSF. After 18 h, cells were pelleted and supernatants of conditioned medium were subjected to SDS-PAGE on 16% Tris-glycine gels under reducing conditions and transferred to polyvinylidene difluoride membrane with detection of immunoreactive bands by polyclonal anti-TIMP-1 Ab. Note that serine protease inhibition reveals the presence of native

35-kDa dog TIMP-1. Downloaded from

CA), respectively. Aliquots removed at specified intervals were centrifuged molecular sizes of ϳ18 and 10 kDa. By contrast, the tryptic mast to pellet cells. To identify the electrophoretic profile of products resulting cell serine protease, tryptase, has no effect in agreement with pre- from processing of native dog TIMP-1, various concentrations of purified viously published data (5). The electrophoretic profile of ␣-chy- dog ␣-chymase (24) were incubated with control BR cell supernatants for different periods of time. Cell supernatants and cleavage reactions were mase-processed TIMP-1 products differs from that of HNE which Ϫ stored at 20°C before analysis by immunoblotting. yields bands at 17 and 16 kDa (30, 31) and additional products of http://www.jimmunol.org/ 8–10 kDa. Thus, these data suggest that ␣-chymase cleavage of Statistical analysis unbound TIMP-1 involves limited hydrolysis to generate products Differences with a p Ͻ 0.05 using Student’s two-tailed t test were consid- that differ from those generated by HNE. As seen in Fig. 2B, in- ered to be significant. cubation of ␣-chymase with TIMP-1 at a 0.5:1 molar ratio yields a faint ϳ18-kDa band. Increasing the molar ratio yields a more Results intense ϳ18-kDa band and an additional band at ϳ10 kDa. The Mastocytoma TIMP-1 expression time course shown in Fig. 2C illustrates that at a ratio of 1:1, Our previous studies established that mast cells express progel B, ␣-chymase generates the ϳ18-kDa band by 5 min with the ϳ10- which could be purified in zymogen form free of TIMP-1 from dog kDa band visible by 15 min. The later appearance of the 10-kDa by guest on September 24, 2021 BR mastocytoma cells, which phenotypically resemble human band suggests that 1) it derives from a separate cleavage site on the mast cells in their protease complement (2, 4, 25). Since cells TIMP-1 parent protein that is less avidly cleaved or 2) that it de- usually secrete pro-MMPs complexed to TIMPs in a 1:1 stoichi- rives from the 18-kDa band via cleavage at a site that is accessible ometric ratio, expression of progel B by mastocytoma cells pre- only after generation of the 18-kDa band. The effect of ␣-chymase dicted coexpression of TIMP-1, which binds the zymogen and ac- processing on the MMP inhibitory activity of TIMP-1 was ex- tive forms of progel B (10, 26). As seen in Fig. 1, A and B, plored by determining its effect on the ability of TIMP-1 to inhibit autoradiography reveals a single TIMP-1 mRNA signal whose soluble activity of active gel B. As shown in Fig. 2D, ␣-chymase steady-state level remains unchanged following KL stimulation of processing of TIMP-1 decreases its ability to inhibit active gel B BR cells; by contrast, KL increases gel B mRNA levels by ϳ5- monomer by Ͼ80% compared with that of unprocessed TIMP-1. fold, as previously described (4). As seen in Fig. 1C, analysis of Thus, cleavage by ␣-chymase attenuates the ability of TIMP-1 to media conditioned by cells incubated for 18 h reveals immunore- inhibit gel B activity in solution. active bands at ϳ12 and 14 kDa (bands b and c), an unexpected ␣ electrophoretic profile since native dog TIMP-1 migrates at 35 -Chymase cleavage sites of TIMP-1 kDa, a size which differs from that of human TIMP-1 (28.5 kDa) To clarify the mechanism of ␣-chymase processing of TIMP-1, we due to the presence of an additional N-linked glycosylation site characterized the electrophoretic profile of the cleavage products (27–29). Since serine proteases process TIMP-1 in vitro (30), cells and identified sites of hydrolysis. As seen in Fig. 3A, analysis of were incubated in the presence of increasing concentrations of cleavage products by SDS-PAGE yields two bands at ϳ18 and 10

PMSF to determine whether proteolysis obscures the identification kDa under reducing conditions. NH2-terminal sequencing of the ϳ of native TIMP-1 bands. Incubation with PMSF yields additional 18-kDa product yields two sequences (X1NX2DLVI and VGT- ϳ8-, 20-, and 27-kDa immunoreactive bands, while the maximum PEV) in a molar ratio of ϳ1:3, while that of the ϳ10-kDa product concentration yields an ϳ35-kDa band (band a). Thus, serine pro- yields two sequences (VAPWN and NSLSLAQ) in a molar ratio of tease inhibition reveals the presence of native dog TIMP-1 protein ϳ1:4, as shown in Fig. 3B. Alignment of these sequences with the and its intermediate cleavage products. These data demonstrate primary sequence of human TIMP-1 (29) reveals that ␣-chymase that mast cells express TIMP-1 mRNA and protein, which may be cleaves the Phe12-Cys13 and Phe23-Val24 bonds in loop 1 and the 101 102 105 106 processed extracellularly by serine proteases released from secre- Phe -Val and Trp -Asn bonds in loop 3 of the NH2-ter- tory granules. minal domain (residues 1–124 (32)). Whereas residue X1 is inde- terminate due to the inability of Edman degradation to detect Cys ␣-Chymase cleaves and inactivates TIMP-1 residues, no residue was assigned to position X2 due to insufficient As shown in Fig. 2A, incubation of rhTIMP-1 with rh␣-chymase at discrimination of chromatographic peak amplitudes in successive a 1:1 molar ratio for 2 h generates two principle products with cycles. Thus, the P1 residue at each of the identified cleavage 2786 ␣-CHYMASE CLEAVAGE OF TIMP-1

FIGURE 3. N-terminal sequences of ␣-chymase-processed TIMP-1 Downloaded from FIGURE 2. ␣-Chymase cleaves and inactivates TIMP-1. A, Purified hu- cleavage products. A, Purified human ␣-chymase and TIMP-1 were incu- man TIMP-1 (25 ␮g/ml) was incubated alone or with purified human bated alone or together at a 0.5:1 molar ratio at 37°C for 30 min. Samples ␣-chymase, a molar equivalent of active tryptase monomer (which exists in were analyzed by SDS-PAGE under reducing conditions with proteins de- tetrameric form in solution) or HNE (25 ␮g/ml) at 37°C for1hinreaction tected by Coomassie blue (A). Cleavage generates two products ϳ18 (band buffer. Reaction products were subjected to SDS-PAGE on 4–20% gradi- a) and 10 kDa (band b) in size which are not visualized in the purified ent gels under reducing conditions and detected by immunoblotting using preparations of either ␣-chymase or TIMP-1. B, Bands a and b were trans- http://www.jimmunol.org/ polyclonal anti-TIMP-1 Ab. B and C, Concentration- and time-dependent ferred to polyvinylidene difluoride membrane, identified by Coomassie ␣-chymase processing of TIMP-1. B, TIMP-1(25 ng) was incubated alone staining, and excised for automated peptide sequencing. Sequences a1 and (C) or with increasing (0.5, 1, 2.5, or 5) molar equivalents of ␣-chymase at a2 (picomoles indicated in parentheses) identify band a as a doublet re- 12 23 37°C for 1 h in reaction buffer. Reaction products were detected by im- sulting from two cleavages at scissile bonds containing Phe and Phe as munoblotting using anti-TIMP-1 Ab. An ϳ18-kDa cleavage product re- the P1 residues. Sequences b1 and b2 also identify band b as a doublet with sulting from incubation with 0.5 molar equivalents of ␣-chymase becomes two overlapping sequences resulting from cleavage at scissile bonds con- 101 105 more intense in the presence of higher concentrations of the enzyme. An taining Phe and Trp . Peptide sequences were aligned with the cDNA ϳ10-kDa product is visible upon incubation with concentrations of ␣-chy- predicted amino acid sequence of human TIMP-1 (29) using MacVector mase at molar equivalence or greater. C, Aliquots of TIMP-1 (25 ␮g/ml) software. incubated with ␣-chymase (25 ␮g/ml) were removed at the indicated time by guest on September 24, 2021 intervals and placed at 0°C in the presence of 2 ␮M AAPF-CMK and reducing buffer. Note the appearance of the 18- and 10-kDa bands by 5 and ␣-Chymase processing of the progel B-TIMP-1-NGAL complex 15 min, respectively. D, ␣-Chymase inactivation of TIMP-1. ␣-Chymase To determine whether ␣-chymase also cleaves bound TIMP-1, in- was incubated alone or with TIMP-1 at a 2.5:1 molar ratio at 37°C for 0 h (uncleaved) or 12 h (cleaved) with the reactions stopped by the addition of creasing concentrations of the enzyme were incubated with a ter- 25 ␮M AAPF-CMK to inhibit ␣-chymase activity. Active gel B monomer nary molecular complex comprised of progel B, TIMP-1, and was incubated with aliquots of each reaction mix at a 1:1 molar ratio to NGAL (progel B-TIMP-1-NGAL). As shown in Fig. 5A, incuba- TIMP-1 for 10 min before determinations of gelatinolytic activity in so- tion of ␣-chymase with the progel B-TIMP-1-NGAL complex at a lution using the fluorogenic substrate DNP-PGCHAK. Values are ex- 0.5:1 molar ratio of ␣-chymase to the TIMP-1 component results pressed as a percentage of inhibition of active gel B monomer relative to in the appearance of an ϳ18-kDa band. Increasing the molar con- that by uncleaved TIMP-1 and represent the mean Ϯ SE of inhibition centration of ␣-chymase yields an additional band at ϳ10 kDa, Ͻ ء analyzed in triplicate ( , p 0.005 compared with uncleaved). while a 10-fold molar excess decreases the 28.5-kDa band of intact TIMP-1. Thus, the electrophoretic profile of cleavage products re- sulting from ␣-chymase processing of bound TIMP-1 is similar to that generated upon cleavage of free TIMP-1. These data suggest sites is aromatic (Phe or Trp) as expected of hydrolysis by a that bound TIMP-1 is susceptible to hydrolysis by ␣-chymase. chymotryptic enzyme. These aromatic residues are completely As seen in Fig. 5B, immunoblot analysis of cleavage products conserved in all mammalian TIMP-1. By contrast, the residue using polyclonal anti-gel B Ab detects bands at 82–84, 42, and on the other (P1Ј) side of the scissile bond varies in mammalian 20–35 kDa. Increasing the molar ratio of ␣-chymase to progel B TIMPs. Met or Leu substitutes for Val24 or Val102 in rodent in the complex from 0.5 to 2.5 decreases the level of detectable sequences, respectively, while His, Arg, and Ser substitute for 92-kDa proenzyme, whereas molar equivalents Ͼ2.5 generate a Asn106 in the rat, mouse, and canine sequences, respectively 42-kDa product. As shown in Fig. 5C, gelatin zymography (per- (Ref. 33 and GenBank Accession no. AAD10632). As seen in formed under nonreducing conditions) detects the disulfide-linked Fig. 4, HNE cleaves TIMP-1 at a single site at the Val69-Cys70 complex of progel B (92 kDa) and NGAL (25 kDa) at ϳ117 kDa bond which is situated in an area of intermolecular contact with following electrophoretic dissociation of TIMP-1, as seen in lane Ͼ stromelysin-1 (MMP-3) (32, 34). By contrast, ␣-chymase inac- C. The band at 203 kDa likely represents progel B homodimer tivates free TIMP-1 by hydrolyzing at four sites in loops 1 and (15). Incubation of the progel B-TIMP-1-NGAL complex in the ␣ 3, which are distant from the active site and located in the presence of increasing molar concentrations of -chymase gener- ates a lower band differing in size by ϳ10–20 kDa. These data five-stranded ␤-barrel of the NH -terminal domain, yielding 2 suggest that ␣-chymase processes TIMP-1-bound progel B via a two identifiable cleavage products. mechanism similar to its activation of progel B monomer, whereby The Journal of Immunology 2787

FIGURE 4. Sites of ␣-chymase cleavage in the NH2-terminal domain of TIMP-1. A, Loop- specific ␣-chymase processing. Purified human ␣ -chymase cleaves the NH2-terminal domain (residues 1–124) at four sites (indicated by ar- rowheads), including Phe12-Cys13 and Phe23- Val24 in loop I and Phe101-Val102 and Trp105- Asn106 in loop III, separating it from the COOH- terminal domain (residues 127–184). Cleavage product containing COOH-terminal domain and portions of NH2-terminal domain with intact di- sulfide linkages (yellow) is indicated in blue. By contrast, HNE cleaves at a single site at the Val69-Cys70 bond (31). B, Ribbon diagram of bound TIMP-1 showing locations of scissile Downloaded from bonds cleaved by ␣-chymase or HNE. Four scis- sile bonds containing aromatic P1 residues fa- vored by ␣-chymase are located away from the long edge of TIMP-1 (red), which contains the Val69-Cys70 bond cleaved by HNE and interacts with the stromelysin (MMP-3) (blue) active site cleft (32, 34). C, Surface localization of aromatic http://www.jimmunol.org/ P1 residues. Scissile bonds containing Phe (green) or Trp (purple) residues at the P1 posi- tion are exposed on the surface of TIMP-1. Po- sitions of the P1 residues are shown in relation to those of Val69 (gray) and Cys70 (yellow) in the bond cleaved by HNE. Figures drawn and ren- dered with SwissPdbViewer v.3.5 based on the crystal structure of the complex formed between

TIMP-1 and the catalytic domain of stromely- by guest on September 24, 2021 sin-1 (MMDB Id: 9217, PDB Id: 1UEA) (34).

removal of an ϳ10-kDa portion of the propeptide domain converts or less (data not shown) of HNE does not increase gelatinolytic the zymogen to active species between ϳ68 and 82 kDa (2, 3). activity of the complex. Therefore, ␣-chymase processing of the progel B-TIMP-1-NGAL complex activates the zymogen, increas- ␣ -Chymase activation of complexed progel B ing the gelatinase activity of the complex in solution. To determine the effect of ␣-chymase processing of TIMP-1- bound progel B, the gel B activity of the progel B-TIMP-1-NGAL Activity of the progel B-TIMP-1-NGAL complex complex in the presence of activator serine proteases was assayed As seen in Fig. 7, the activity of progel B-TIMP-1-NGAL in so- in solution over 2 h using the fluorogenic substrate DNP- lution after processing by activator proteases was compared with PGCHAK. As seen in Fig. 6, incubation of ␣-chymase with the that of a molar equivalent of active gel B monomer. Whereas complex at a 0.1:1 molar ratio increases activity 2-fold. Activity of ␣-chymase increases the activity of the progel B-TIMP-1-NGAL the complex alone likely occurs due to contaminating amounts of complex, neither HNE alone nor a 4-h pretreatment of the complex gel B monomer (35). Coincubation of the complex with ␣-chy- with HNE before addition of 0.2 or 0.5 molar equivalents of active mase at ratios of 0.5:1 or 1:1 increases activity by 3.5- or 4.5-fold, MMP-3 (a mechanism of sequential proteolytic processing previ- respectively. Incubation of molar excesses of ␣-chymase with the ously shown to activate the progel B-TIMP-1 complex (30)) has complex does not further increase activity (data not shown). Sim- any effect. Maximal soluble activity of ␣-chymase-processed com- ilar molar ratios of trypsin, an activator of progel B monomer (35) plex represents ϳ80% of the maximal activity of a molar equiv- (and additional log ratios between 10Ϫ10 and 10Ϫ1; data not alent of an active 82-kDa TIMP-free gel B monomer (relative to shown), do not increase activity to the magnitude induced by the 92-kDa progel B moiety in the complex). Thus, these data ␣-chymase processing. These data demonstrate that the observed suggest that processing of the progel B-TIMP-1-NGAL complex increases in the complex’s soluble activity following ␣-chymase by ␣-chymase overcomes TIMP-1 inhibition and results in recov- processing do not result from its activation of contaminating pro- ery of the soluble activity of the gel B moiety following ␣-chy- gel B monomer. Incubation of the complex with a molar equivalent mase-dependent zymogen activation. 2788 ␣-CHYMASE CLEAVAGE OF TIMP-1

FIGURE 6. ␣-Chymase processing increases soluble activity of the pro- gel B-TIMP-1-NGAL complex. The progel B-TIMP-1-NGAL complex ␣ Downloaded from FIGURE 5. -Chymase cleavage of the progel B-TIMP-1-NGAL com- (145 ng) was incubated in the absence or presence of purified human plex. Purified human progel B-TIMP-1-NGAL complex (145 ng) was in- ␣-chymase, HNE, or trypsin at 37°C for2hinreaction buffer containing cubated alone (C) or in the presence of the indicated molar equivalents of excess fluorogenic substrate, DNP-PGCHAK. Activator proteases were in- ␣ purified human -chymase at 37°C for1hinreaction buffer. Reaction cubated with the complex in ratios relative to the TIMP-1 moiety at 0, 0.1, products were subjected either to immunoblotting with antisera to TIMP-1 0.5, and 1 molar equivalents. Soluble gel B activity was determined in the (A)orgelB(B) after SDS-PAGE under reducing conditions or to gelatin presence of an excess of the fluorogenic substrate DNP-PGCHAK. Values ء zymography performed under nonreducing conditions (C). Ϯ represent the mean SE of gelatinolytic activity analyzed in triplicate ( , http://www.jimmunol.org/ p ϭ 0.0001 compared with complex ,ءءء ;p Ͻ 0.0001 ,ءء ;p Ͻ 0.002 alone). Degranulation-dependent mast cell TIMP-1 processing To determine the effect of degranulation on mast cell TIMP-1 pro- changed, suggesting that its presence in supernatants is not due to cessing, cells were incubated alone or in the presence of 2 ␮M extracellular proteolytic processing. Furthermore, the band appears calcium ionophore A23187 for 5 min. As seen in Fig. 8A, super- not only in the supernatants of BR and HMC-1 cells (which are natant of control dog BR mastocytoma cells contains an immuno- derived from a malignant phenotype), but also in that of MBMMC reactive band at 35 kDa (band a), suggesting constitutive secretion

cells, which suggests that normal mast cells also release a 14-kDa by guest on September 24, 2021 of native TIMP-1; an additional band at 14 kDa (band b) suggests form of TIMP-1 in a constitutive fashion. As seen in Fig. 8C, synchronous release of a related TIMP-1 product. Supernatant of either HMC-1 cells, a human mast cell leukemia line (20), or MBMMC reveals a similar pattern with bands migrating at 30 or 31 kDa, the expected sizes of human (with glycosylation-depen- dent variation from 28.5 to 34 kDa) or murine TIMP-1, respec- tively (29, 36, 37)). Supernatant of HMC-1 or MBMMC cells also contains the 14-kDa immunoreactive TIMP-1 band. These data suggest that mast cells constitutively release native TIMP-1 and that TIMP-1 secretion occurs in cells with either a normal or ma- lignant phenotype. Upon degranulation, the electrophoretic profile of the supernatants of BR or MBMMC cells changes with the appearance of more bands between ϳ17 and 29 kDa, in addition to the native and 14-kDa TIMP-1 bands present in control superna- tants. By contrast, the profile of TIMP-1 bands present in HMC-1 degranulation supernatants remains identical to that of the control. HMC-1 cells are similar to BR or MBMMC cells in their expres- sion of tryptase, which does not cleave TIMP-1 (as seen in Fig. 1), but differ in their lack of expression of ␣-chymase (38). Therefore, FIGURE 7. Relative gel B activity of ␣-chymase-processed progel these data suggest that degranulated mast cell ␣-chymase cleaves B-TIMP-1-NGAL complex. Seventy-three nanograms of the progel TIMP-1 extracellularly. B-TIMP-1-NGAL complex (46 ng of 92-kDa progel B moiety) were in- To identify the protease(s) involved in degranulation-associated cubated alone or in the presence of either purified human ␣-chymase or TIMP-1 processing, BR cells were incubated with calcium iono- HNE without or with active stromelysin (MMP-3). ␣-Chymase or HNE phore A23187 alone or in combination with individual class-spe- was incubated alone with the complex at ratios of molar equivalence with cific inhibitors. As seen in Fig. 8B, incubation of cells with PMSF the TIMP-1 moiety at 37°C for 8 h, while 0.2 or 0.5 molar equivalents of active stromelysin were added after prior treatment of the complex with blocks the appearance of band c while aprotinin has no effect, thus HNE for 4 h. Reactions were performed in the presence of an excess of the implicating nontryptic serine proteases in processing of endoge- fluorogenic substrate DNP-PGCHAK. Soluble gel B activity of the com- nous TIMP-1. By contrast, E64, phosphoramidon, or BB94 does plex alone or in the presence of activator proteases was compared with that not alter the profile, suggesting that cysteine proteases and metal- of 41 ng of active 82-kDa gel B monomer. Values represent the mean Ϯ p Ͻ 0.001 compared ,ء) loproteinases are not involved in extracellular TIMP-1 processing. SE of gelatinolytic activity analyzed in triplicate By contrast, the appearance of the 14-kDa band remains un- with the complex alone or with HNE with or without stromelysin). The Journal of Immunology 2789 Downloaded from

FIGURE 8. Effects of degranulation and purified ␣-chymase on endogenous mast cell TIMP-1. A, BR dog mastocytoma cells, HMC-1 cells, or MBMMC were incubated alone or degranulated using 2 ␮M calcium ionophore A23187 for 5 min at 37°C. Supernatants were subjected to SDS-PAGE on 16% gels under reducing conditions and immunoreactive bands were detected using polyclonal anti-TIMP-1 Ab. Note the absence of TIMP-1 processing in ␣-chy- mase-deficient HMC-1 cells and the constitutive expression of both native TIMP-1 and the 14-kDa form in these mast cells. B, To identify proteases http://www.jimmunol.org/ involved in degranulation-dependent processing of endogenous TIMP-1, BR cells were incubated in the absence (Ϫ) or presence of ionophore, both alone (ϩ) and in combination with molar excesses of class-specific inhibitors: 10 mM PMSF, 10 ␮M aprotinin, 250 ␮M E64, 250 ␮g/ml phosphoramidon, or 500 nM BB94. Note attenuation of TIMP-1 processing by PMSF with the absence of band c, suggesting inhibition of serine protease-dependent processing of endogenous TIMP-1. C, To identify the specific mast cell serine protease involved in degranulation-dependent TIMP-1 processing, BR cells were incubated alone (Ϫ), with ionophore alone (ϩ), and in combination with increasing concentrations (200 nM, 2 ␮M, 20 ␮M, 200 ␮M, 400 ␮M, and 2 mM) of AAPF-CMK, a specific inhibitor of ␣-chymase, for 15 min at 37°C before analysis of supernatants by immunoblotting with polyclonal anti-TIMP-1 Ab. Note the diminution in the intensity of band c and that of intermediate TIMP-1 products in the presence of increasing concentrations of AAPF. D, To demonstrate that ␣-chymase cleaves endogenous TIMP-1, 50 ng of purified dog ␣-chymase was incubated for 30 or 60 min with cell-free supernatants of medium conditioned by BR cells incubated alone for 5 min at 37°C. Exogenous ␣-chymase generates additional ϳ8-, 12-, 17-, and 27-kDa bands, which are also seen in conditioned medium (CM) of BR cells incubated for 18 h at 37°C in the presence of 2 mM PMSF (as shown in Fig. 1C). by guest on September 24, 2021 incubation of cells with ionophore and increasing concentrations cleaving the catalytic domain (3). The work here demonstrates that of AAPF-CMK (a serine protease inhibitor which inhibits ␣-chy- ␣-chymase also cleaves and inactivates TIMP-1 whether it exists mase, but not tryptase) (2), diminishes the intensity of band c and free or bound in a complex with progel B. By cleaving TIMP-1 at blocks the appearance of the 17- to 29-kDa bands induced by de- exposed surface sites, ␣-chymase overcomes its inhibitory effects granulation. As shown in Fig. 8D, incubation of purified dog and converts complexed progel B to mature species with gelatino- ␣-chymase with control BR cell supernatants generates ϳ8-, 12 lytic activity in solution. (band c)-, 17-, and 27-kDa bands which share an electrophoretic Whereas previous work demonstrated that mast cells express profile similar to that of media conditioned by BR cells for 18 h (as MMPs (2, 4, 12, 13, 25), the current data provide evidence that seen in Fig. 1); differences in the molecular weight of TIMP-1 HMC-1 cells, dog BR mastocytoma cells, and MBMMC also se- cleavage products may result from species-dependent variations in crete TIMP-1. Our previous data showed that ligation of mast cell glycosylation. Thus, addition of ␣-chymase alone results in pro- kit receptor by KL up-regulates progel B mRNA expression ϳ5- cessing of TIMP-1 present in BR cell supernatants. These data fold. By contrast, TIMP-1 mRNA levels remain unchanged fol- identify ␣-chymase as the degranulated mast cell serine protease lowing KL stimulation, possibly due to distinct mechanisms which responsible for extracellular cleavage of TIMP-1. regulate TIMP-1 expression via enhancement of mRNA stability Discussion (41). Thus, KL-kit interactions may alter local progel B to TIMP-1 ratios, favoring release of TIMP-1-free progel B, which would be TIMPs act in diverse biological processes via mechanisms which more susceptible to activation. Mast cells not only release the na- may be either dependent on or independent of their ability to in- ϳ hibit MMPs. In part, their myriad effects may be due to interactions tive form of TIMP-1, but also an 14-kDa form whose presence in conditioned medium is explained neither by extracellular pro- involving structurally distinct NH2- and COOH-terminal domains, which block MMP-active sites and bind to MMP COOH domains, teolytic processing nor by translation of alternatively spliced respectively (8). However, regulation of these TIMP domains is mRNA. Whether posttranslational intracellular processing or non- unclear. A unique feature of the gelatinase class of MMPs is bind- proteolytic extracellular mechanisms generate the 14-kDa form re- ing of the zymogens, progelatinase A (MMP-2) and progel B, to mains a subject of investigation. Both the native and ϳ14-kDa TIMP-2 and TIMP-1, respectively (39, 40). Thus, TIMP-1 not only forms of TIMP-1 also appear to be products of normal MBMMC, inhibits activity of mature gel B, but also blocks activation of the thus suggesting that the secreted TIMP-1 products of cultured ca- proenzyme. We previously reported that ␣-chymase, a chymotryp- nine and human mast cells do not only represent a malignant tic serine protease unique to mast cells, activates progel B by phenotype. 2790 ␣-CHYMASE CLEAVAGE OF TIMP-1

␣-Chymase inactivates TIMP-1 by hydrolyzing bonds in the cleaves TIMP-1 at multiple sites in loops 1 and 3, which may

NH2-terminal domain. Whereas cleavage by HNE, trypsin, or chy- separate the NH2- and COOH-terminal domains. This contrasts motrypsin destroys the inhibitory activity of unbound TIMP-1 with the single cleavage of TIMP-1 by HNE in loop 1, which (31), only HNE has previously been shown to degrade TIMP-1 in leaves the domains linked via the disulfide bonds. Perhaps the ␣ ␣ the progel B-TIMP-1 complex (30). Like HNE, -chymase cleaves liberated, inactivated NH2-terminal domain arising from -chy- TIMP-1, whether the inhibitor exists free or in a complex with mase cleavage dissociates from the complex, facilitating subse- progel B. However, ␣-chymase generates products that differ in quent activation of progel B. Whether the TIMP-1 COOH domain size from those generated by HNE, suggesting that hydrolysis by retains its non-MMP inhibitory functions following ␣-chymase ␣ -chymase occurs at different sites. The four scissile bonds iden- processing remains unknown. tified all contain aromatic P1 residues (Phe or Trp), which are Hyperplasia and degranulation are key features of the mast cell ␣ favored by -chymase and highly conserved in mammalian spe- response in inflamed tissues. Degranulating mast cells release cies, as seen in Fig. 4; peptide sequencing did not identify cleavage ␣-chymase in active form from secretory granules where it is ␣ at sites containing other -chymase-preferred P1 residues (includ- stored in high concentrations, second only in abundance to tryptase ing Tyr and a hydrophobic residue, Leu), which localize to the (1, 47). The current data demonstrate that secreted mast cell surface of TIMP-1 (34, 42). This suggests that cleavage at these TIMP-1 remains in its native form until ionophore-induced de- sites may not be favorable due to steric hindrance or that the cleav- granulation releases ␣-chymase, which cleaves the inhibitor to age products may either be transient or insufficient in quantity for generate products similar to those following incubation of purified detection and sequencing. Therefore, although ␣-chymase and ␣-chymase with mast cell supernatants. Lack of TIMP-1 process- HNE are similar in their inactivation of free or bound TIMP-1 via Downloaded from ing by degranulating HMC-1 cells, which do not express ␣-chy- NH2-terminal domain processing, they differ in the number as well as general location of TIMP-1 bonds cleaved. mase, further substantiates the dependence of mast cell TIMP-1 ␣ Conversion of complexed progel B to active forms requires pro- processing on exocytosed -chymase. Whether nonproteolytic cessing which overcomes zymogen latency conferred by the cys- mechanisms may also process secreted TIMP-1 remains unknown. ␣ ϳ teine switch (9) and inhibition of active species by bound TIMP-1. Mast cells express -chymase at levels of 5 pg/cell (48), yielding a cellular concentration of ϳ300 ␮M (which underestimates the Cells secrete pro-MMPs noncovalently bound to TIMPs with 1:1 http://www.jimmunol.org/ stoichiometry, suggesting that physiologic regulation of MMP ac- concentration in secretory granules due to differences in granule tivity depends on one or more activators, acting alone or in con- and cytosolic volumes) that predicts even higher local tissue con- cert, to activate secreted proenzymes. Proteases which hydrolyze centrations upon degranulation. In human tissues, ␣-chymase is either free TIMP-1 or uncomplexed progel B monomer in vitro found at a concentration of as much as ϳ45 ␮g/g of tissue (48, 49), cannot cleave both components when complexed to each other (17, compared with a TIMP-1 concentration of ϳ12.4 ␮g/g (50). Im- 30, 31). By contrast, at catalytic or stoichiometric molar ratios, munolocalization studies in a murine model of squamous epithelial ␣-chymase processing of the progel B-TIMP-1-NGAL complex carcinogenesis demonstrate limited diffusability of chymase; thus, results not only in cleavage of both TIMP-1 and progel B, but also it remains in the vicinity of degranulating mast cells situated near in conversion of the zymogen to mature forms with up to a 5-fold basement membranes, which are sites of extracellular matrix re- by guest on September 24, 2021 increase in soluble activity. Activation of the zymogen in the pro- modeling (11). The abundance of stored ␣-chymase and its limited gel B-TIMP-1-NGAL complex also suggests that NGAL (bound diffusability suggest that in vivo mast cell degranulation may gen- via disulfide linkage to progel B without altering the protease’s erate transient and localized molar excesses of ␣-chymase relative activity (15, 43, 44)) does not hinder access of ␣-chymase to either to its substrates such as TIMP-1. We speculate that such a quantum TIMP-1 or progel B. Processing by catalytic amounts of ␣-chy- proteolytic event (51) may be an important mechanism whereby mase alone compares favorably with other proteolytic mechanisms exocytosed mast cell ␣-chymase regulates both the MMP inhibi- which activate the complex via 1) molar excesses of either active tory and noninhibitory functions of TIMP-1 in remodeling tissues. stromelysin or matrilysin (MMP-7), which increase gelatinolytic Evidence of ongoing and gradual (or piecemeal) degranulation of activity by competitively binding TIMP-1 (without cleavage) and mast cells in fibrotic tissues (52) suggests that such a mechanism activating progel B (45, 46); or 2) sequential involvement of HNE, may not only occur transiently, but also persist indefinitely. There- which cleaves and inactivates TIMP-1, and mature stromelysin, fore, the quiescence or activation of mast cells in lung, skin, heart, which activates progel B (30). In the present work, combined HNE or gut tissues may determine not only the protease to antiprotease and stromelysin processing of the complex did not increase its gel balance of MMPs to TIMPs which is critical in matrix remodeling B activity, possibly due to differences in substrates and reaction and fibrogenesis (8, 52, 53), but also the ability of TIMP-1 to conditions. Thus, ␣-chymase processing of the progel B-TIMP-1- regulate apoptosis and mitogenesis via pathways which are inde- NGAL complex not only cleaves and inactivates bound TIMP-1, pendent of its MMP inhibitory functions (11, 54). whose hydrolysis appears to be rate limiting, but also converts ␣ progel B from zymogen to active forms. The current data suggest a role for mast cell -chymase in phys- iologic pathways which regulate MMP activity. Unlike other pro- Access to the NH2-terminal domain of complexed TIMP-1 is critical for cleavage and attenuation of its inhibition of gel B. In- teases which require cascading activation of multiple proenzymes activation of TIMP-1 by HNE depends on cleavage of the Val69- before activation of the target pro-MMP (2, 3, 5, 8, 17, 35, 55–57), ␣ Cys70 bond in the “C-connector loop” which forms the long edge -chymase activates zymogen monomers of progel B, procollage- that occupies the MMP-active site cleft (34). This region remains nase, and stromelysin, without a requirement for other proteases or exposed and accessible to HNE when TIMP-1 binds to progel B. cofactors (2, 3, 6, 7). Data also show that infiltration of premalig- Other serine proteases such as trypsin and can nant lesions by mast cells activates progel B via ␣-chymase pro- cleave free TIMP-1, but cannot process bound TIMP-1, suggesting cessing, coincident with the activation of angiogenesis, thus dem- that their preferred scissile bonds are obscured in the complex (30). onstrating an in vivo role for ␣-chymase-dependent pro-MMP By contrast, TIMP-1 regions cleaved by ␣-chymase are distant activation in tissue remodeling (11). In summary, our data show from the long edge and are thus likely to be accessible whether the that cultured mast cells, such as HMC-1 or BR cells and MBMMC, inhibitor is free or bound, as seen in Fig. 4. ␣-Chymase also secrete TIMP-1, which is processed by exocytosed ␣-chymase. The Journal of Immunology 2791

Uniquely, ␣-chymase processes the progel B-TIMP-1-NGAL ter- 20. Butterfield, J. H., D. Weiler, G. Dewald, and G. J. Gleich. 1988. Establishment nary complex by inactivating bound TIMP-1 and converting pro- of an immature mast cell line from a patient with mast cell leukemia. Leuk. Res. 12:345. gel B to mature active forms. ␣-Chymase hydrolysis of TIMP-1 21. Razin, E., J. N. Ihle, D. Seldin, J. M. Mencia-Huerta, H. R. Katz, P. A. LeBlanc, occurs at four scissile bonds containing exposed aromatic residues A. Hein, J. P. Caulfield, K. F. Austen, and R. L. Stevens. 1984. Interleukin 3: a differentiation and growth factor for the mouse mast cell that contains chondroitin distant from regions which interact with the MMP active site. sulfate E proteoglycan. J. Immunol. 132:1479. Cleavage attenuates the inhibitory activity of TIMP-1 and may 22. Caughey, G. H., W. W. Raymond, and P. J. Wolters. 2000. Angiotensin II gen- eration by mast cell ␣- and ␤-chymases. Biochim. Biophys. Acta. 1480:247. separate its NH2- and COOH-terminal domains. Therefore, release of ␣-chymase may provide a novel mechanism for recovering and 23. Bickett, D. M., M. D. Green, J. Berman, M. Dezube, A. S. Howe, P. J. Brown, J. T. Roth, and G. M. McGeehan. 1993. A high throughput fluorogenic substrate augmenting activity of progel B complexed to TIMP-1 during ex- for interstitial collagenase (MMP-1) and gelatinase (MMP-9). Anal. Biochem. tracellular matrix remodeling in mast cell-rich environments. 212:58. 24. Caughey, G. H., N. F. Viro, S. C. Lazarus, and J. A. Nadel. 1988. Purification and characterization of dog mastocytoma chymase: identification of an octapeptide Acknowledgments conserved in chymotryptic leukocyte proteinases. Biochim. Biophys. Acta 952: 142. We thank Ralph Reid and Wilfred W. Raymond for helpful discussions and 25. Kanbe, N., A. Tanaka, M. Kanbe, A. Itakura, M. Kurosawa, and H. Matsuda. Paul J. Wolters for providing murine mast cells. We also acknowledge the 1999. Human mast cells produce matrix metalloproteinase 9. Eur. J. Immunol. excellent technical assistance of Afshin Bidgol, John Blount, and Diego 29:2645. Muilenburg. 26. Olson, M. W., D. C. Gervasi, S. Mobashery, and R. Fridman. 1997. Kinetic analysis of the binding of human matrix metalloproteinase-2 and -9 to tissue inhibitor of metalloproteinase (TIMP)-1 and TIMP-2. J. Biol. Chem. 272:29975. References 27. Chopra, R., P. A. Koklitis, S. Bergin, J. Rowe, and S. Angal. 1991. Purification 1. Caughey, G. H. 1991. The structure and airway biology of mast cell proteinases. of recombinant dog tissue inhibitor of metalloproteinases. Biochem. Soc. Trans. Downloaded from Am. J. Respir. Cell Mol. Biol. 4:387. 19:372S. 2. Fang, K. C., W. W. Raymond, S. C. Lazarus, and G. H. Caughey. 1996. Dog 28. Zeiss, C. J., G. M. Acland, G. D. Aguirre, and K. Ray. 1998. TIMP-1 expression mastocytoma cells secrete a 92-kD gelatinase activated extracellularly by mast is increased in X-linked progressive retinal atrophy despite its exclusion as a cell chymase. J. Clin. Invest. 97:1589. candidate gene. Gene 225:67. 3. Fang, K. C., J. Blount, W. W. Raymond, and G. H. Caughey. 1997. Dog mast cell 29. Docherty, A. J., A. Lyons, B. J. Smith, E. M. Wright, P. E. Stephens, T. J. Harris, ␣-chymase activates progelatinase B by cleaving the Phe88-Gln89 and Phe91- G. Murphy, and J. J. Reynolds. 1985. Sequence of human tissue inhibitor of Glu92 bonds of the catalytic domain. J. Biol. Chem. 272:25628. metalloproteinases and its identity to erythroid-potentiating activity. Nature 318:

4. Fang, K. C., P. J. Wolters, M. Steinhoff, A. Bidgol, J. L. Blount, and 66. http://www.jimmunol.org/ G. H. Caughey. 1999. Mast cell expression of gelatinases A (MMP-2) and B 30. Itoh, Y., and H. Nagase. 1995. Preferential inactivation of tissue inhibitor of (MMP-9) is regulated by kit ligand and transforming growth factor-␤. J. Immu- metalloproteinases-1 that is bound to the precursor of matrix metalloproteinase 9 nol. 162:5528. (progelatinase B) by human neutrophil elastase. J. Biol. Chem. 270:16518. 5. Gruber, B. L., M. J. Marchese, K. Suzuki, L. B. Schwartz, Y. Okada, H. Nagase, 31. Okada, Y., S. Watanabe, I. Nakanishi, J. Kishi, T. Hayakawa, W. Watorek, and N. S. Ramamurthy. 1989. Synovial procollagenase activation by human mast J. Travis, and H. Nagase. 1988. Inactivation of tissue inhibitor of metallopro- cell tryptase dependence upon matrix metalloproteinase 3 activation. J. Clin. teinases by neutrophil elastase and other serine proteinases. FEBS Lett. 229:157. Invest. 84:1657. 32. Nagase, H., K. Suzuki, T. E. Cawston, and K. Brew. 1997. Involvement of a 6. Saarinen, J., N. Kalkkinen, H. G. Welgus, and P. T. Kovanen. 1994. Activation region near valine-69 of tissue inhibitor of metalloproteinases (TIMP)-1 in the of human interstitial procollagenase through direct cleavage of the Leu83-Thr84 interaction with matrix metalloproteinase 3 (stromelysin 1). Biochem. J. 325:163. bond by mast cell chymase. J. Biol. Chem. 269:18134. 33. Douglas, D. A., Y. E. Shi, and Q. A. Sang. 1997. Computational sequence anal- 7. Suzuki, K., M. Lees, G. F. J. Newlands, H. Nagase, and D. E. Woolley. 1995. ysis of the tissue inhibitor of metalloproteinase family. J. Protein Chem. 16:237. Activation of precursors for matrix metalloproteinases 1 (interstitial collagenase) 34. Gomis-Ruth, F. X., K. Maskos, M. Betz, A. Bergner, R. Huber, K. Suzuki, by guest on September 24, 2021 and 3 (stromelysin) by rat mast-cell proteinases I and II. Biochem. J. 305:301. N. Yoshida, H. Nagase, K. Brew, G. P. Bourenkov, et al. 1997. Mechanism of 8. Nagase, H., and J. F. Woessner, Jr. 1999. Matrix metalloproteinases. J. Biol. inhibition of the human matrix metalloproteinase stromelysin-1 by TIMP-1. Na- Chem. 274:21491. ture 389:77. 9. Springman, E. B., E. L. Angleton, H. Birkedal-Hansen, and H. E. Van Wart. 35. Ogata, Y., Y. Itoh. and H. Nagase. 1995. Steps involved in activation of the 1990. Multiple modes of activation of latent human fibroblast collagenase: evi- 73 pro-matrix metalloproteinase 9 (progelatinase B)-tissue inhibitor of metallopro- dence for the role of a Cys active-site zinc complex in latency and a “cysteine teinases-1 complex by 4-aminophenylmercuric acetate and proteinases. J. Biol. switch” mechanism for activation. Proc. Natl. Acad. Sci. USA 87:364. Chem. 270:18506. 10. Brew, K., D. Dinakarpandian, and H. Nagase. 2000. Tissue inhibitors of metal- 36. Rosenthal, R. A., M. A. Moses, Y. Shintani, J. F. Megyesi, R. Langer, and loproteinases: evolution, structure and function. Biochim. Biophys Acta 1477: J. Folkman. 1994. Purification and characterization of two collagenase inhibitors 267. from mouse sarcoma 180 conditioned medium. J. Cell. Biochem. 56:97. 11. Coussens, L. M., W. W. Raymond, G. Bergers, M. Laig-Webster, 37. Gomez, D. E., D. F. Alonso, H. Yoshiji, and U. P. Thorgeirsson. 1997. Tissue O. Behrendtsen, Z. Werb, G. H. Caughey, and D. Hanahan. 1999. Inflammatory inhibitors of metalloproteinases: structure, regulation and biological functions. mast cells up-regulate angiogenesis during squamous epithelial carcinogenesis. Eur. J. Cell Biol. 74:111. Genes Dev. 13:1382. 38. Xia, H. Z., C. L. Kepley, K. Sakai, J. Chelliah, A. M. Irani, and L. B. Schwartz. 12. Tozzi, C. A., S. Thakker-Varia, S. Y. Yu, R. F. Bannett, B. W. Peng, G. J. Poiani, ⑀ ␣ ⑀ ␥ F. J. Wilson, and D. J. Riley. 1998. Mast cell collagenase correlates with regres- 1995. Quantitation of tryptase, chymase, Fc RI , and Fc RI mRNAs in human sion of pulmonary vascular remodeling in the rat. Am. J. Respir. Cell Mol. Biol. mast cells and by competitive reverse transcription-polymerase chain 18:497. reaction. J. Immunol. 154:5472. 13. Brownell, E., L. Fiorentino, G. Jolly, K. Wolfe, S. Kincaid, P. Seperack, and 39. Wilhelm, S. M., I. E. Collier, B. L. Marmer, A. Z. Eisen, G. A. Grant, and D. Visco. 1995. Immunolocalization of stromelysin-related protein in murine G. I. Goldberg. 1989. SV40-transformed human lung fibroblasts secrete a 92-kDa mast cell granules. Int. Arch. Allergy Immunol. 107:333. type IV collagenase which is identical to that secreted by normal human macro- 14. Butler, G. S., S. S. Apte, F. Willenbrock, and G. Murphy. 1999. Human tissue phages. J. Biol. Chem. 264:17213. inhibitor of metalloproteinases 3 interacts with both the N- and C-terminal do- 40. Goldberg, G. I., B. L. Marmer, G. A. Grant, A. Z. Eisen, S. Wilhelm, and mains of gelatinases A and B: regulation by polyanions. J. Biol. Chem. 274: C. S. He. 1989. Human 72-kilodalton type IV collagenase forms a complex with 10846. a tissue inhibitor of metalloproteases designated TIMP-2. Proc. Natl. Acad. Sci. 15. Kolkenbrock, H., A. Hecker-Kia, D. Orgel, A. Kinawi, and N. Ulbrich. 1996. USA 86:8207. Progelatinase B forms from human neutrophils. complex formation of monomer/ 41. Doyle, G. A., U. K. Saarialho-Kere, and W. C. Parks. 1997. Distinct mechanisms lipocalin with TIMP-1. Biol. Chem. 377:529. regulate TIMP-1 expression at different stages of phorbol ester-mediated differ- 16. Frears, E. R., Z. Zhang, D. R. Blake, J. P. O’Connell, and P. G. Winyard. 1996. entiation of U937 cells. Biochemistry 36:2492. Inactivation of tissue inhibitor of metalloproteinase-1 by peroxynitrite. FEBS 42. Powers, J. C., T. Tanaka, J. W. Harper, Y. Minematsu, L. Barker, D. Lincoln, Lett. 381:21. K. V. Crumley, J. E. Fraki, N. M. Schechter, G. G. Lazarus, et al. 1985. Mam- 17. Ogata, Y., J. J. Enghild, and H. Nagase. 1992. Matrix metalloproteinase 3 malian chymotrypsin-like : comparative reactivities of rat mast cell pro- (stromelysin) activates the precursor for the human matrix metalloproteinase 9. teases, human and dog skin chymases, and human G with peptide J. Biol. Chem. 267:3581. 4-nitroanilide substrates and with peptide chloromethyl ketone and sulfonyl fluo- 18. Okada, Y., Y. Gonoji, K. Naka, K. Tomita, I. Nakanishi, K. Iwata, K. Yamashita, ride inhibitors. Biochemistry 24:2048. and T. Hayakawa. 1992. Matrix metalloproteinase 9 (92-kDa gelatinase/type IV 43. Triebel, S., J. Blaser, H. Reinke, V. Knauper, and H. Tschesche. 1992. Mercurial collagenase) from HT 1080 human fibrosarcoma cells: purification and activation activation of human PMN leucocyte type IV procollagenase (gelatinase). FEBS of the precursor and enzymic properties. J. Biol. Chem. 267:21712. Lett. 298:280. 19. Kostoulas, G., A. Lang, H. Nagase, and A. Baici. 1999. Stimulation of angio- 44. Coles, M., T. Diercks, B. Muehlenweg, S. Bartsch, V. Zolzer, H. Tschesche, and genesis through cathepsin B inactivation of the tissue inhibitors of matrix met- H. Kessler. 1999. The solution structure and dynamics of human neutrophil ge- alloproteinases. FEBS Lett. 455:286. latinase- associated lipocalin. J. Mol. Biol. 289:139. 2792 ␣-CHYMASE CLEAVAGE OF TIMP-1

45. Birkedal-Hansen, H., W. G. Moore, M. K. Bodden, L. J. Windsor, 51. Campbell, E. J., M. A. Campbell, S. S. Boukedes, and C. A. Owen. 1999. Quan- B. Birkedal-Hansen, A. DeCarlo, and J. A. Engler. 1993. Matrix metalloprotein- tum proteolysis by neutrophils: implications for pulmonary emphysema in ␣1- ases: a review. Crit. Rev. Oral Biol. Med. 4:197. antitrypsin deficiency. J. Clin. Invest. 104:337. 46. von Bredow, D. C., A. E. Cress, E. W. Howard, G. T. Bowden, and R. B. Nagle. 52. Kawanami, O., V. J. Ferrans, J. D. Fulmer, and R. G. Crystal. 1979. Ultrastruc- 1998. Activation of gelatinase-tissue-inhibitors-of-metalloproteinase complexes ture of pulmonary mast cells in patients with fibrotic lung disorders. Lab. Invest. 40:717. by matrilysin. Biochem. J. 331:965. 53. Schulman, E. S. 1993. The role of mast cells in inflammatory responses in the 47. Schwartz, L. B., A. M. Irani, K. Roller, M. C. Castells, and N. M. Schechter. lung. Crit. Rev. Immunol. 13:35. 1987. Quantitation of histamine, tryptase, and chymase in dispersed human T and 54. Jeziorska, M., C. McCollum, and D. E. Woolley. 1997. Mast cell distribution, TC mast cells. J. Immunol. 138:2611. activation, and phenotype in atherosclerotic lesions of human carotid arteries. 48. Caughey, G. H. 1995. Chymase. In Human Protein Data. A. Haeberli, ed. VCH J. Pathol. 182:115. Verlgasges, Weinheim. 55. Desrivieres, S., H. Lu, N. Peyri, C. Soria, Y. Legrand, and S. Menashi. 1993. Activation of the 92 kDa type IV collagenase by tissue . J. Cell. Physiol. 49. Matin, R., E. K. Tam, J. A. Nadel, and G. H. Caughey. 1992. Distribution of 157:587. chymase-containing mast cells in human bronchi. J. Histochem. Cytochem. 40: 56. Fridman, R., M. Toth, D. Pena, and S. Mobashery. 1995. Activation of progela- 781. tinase B (MMP-9) by gelatinase A (MMP-2). Cancer Res. 55:2548. 50. Jung, K., M. Lein, N. Ulbrich, B. Rudolph, W. Henke, D. Schnorr, and 57. Ramos-DeSimone, N., E. Hahn-Dantona, J. Sipley, H. Nagase, D. L. French, and S. A. Loening. 1998. Quantification of matrix metalloproteinases and tissue in- J. P. Quigley. 1999. Activation of matrix metalloproteinase-9 (MMP-9) via a hibitors of metalloproteinase in prostatic tissue: analytical aspects. Prostate 34: converging /stromelysin-1 cascade enhances tumor cell invasion. J. Biol. 130. Chem. 274:13066. Downloaded from http://www.jimmunol.org/ by guest on September 24, 2021