THE JOURNAL OF BIOWGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, lnc.

Vol. 269, No. 51, Issue of December 23, pp. 32660-32666, 1994 Printed in U.S.A.

Regulation of Macrophage Receptor-bound Plasmin by Autoproteolysis"

(Received for publication, October 4, 1994)

Domenick J. FalconelOn, Wolfgang Borthll, Timothy A. McCaffrey**, Jean Mathew$, and Kimberly McAdaml From the Departments of $.Pathology, $Cell Biology and Anatomy, and **Medicine, Cornell Medical College and the JDiuision of Hematology, Mt. Sinai School of Medicine, New York, New York 10021

The activation of plasminogen by macrophage is regu- lated by their expression of receptors for urokinase and plasmin(ogen). In these studies we have examined plas- midogen) binding to adherent human THP-1 macro- phage. Plasmin bound to the THP-1 cells in a time- and dose-dependent manner (K, 15.8 t 6.2 m; B , 1.4 * 0.3 x 106/cell). The lysine analog e-aminocaproic acid competi- tively inhibited plasmin binding. The fraction of mem- brane-bound plasmin, however, became increasingly re- sistant to displacement with eaminocaproic acid. Over a 24-h period, membrane-bound plasmin activity fell 80% despite the presence of catalytically active plasmin in the incubation media. The loss of receptor-bound plas- min activity was not due to proteolytic alterations of its receptor since '2SI-Lys-plasminogen bound to THP-1 cells pretreated with plasmin with similar affinity as to un- treated cells. Following a 24-h incubation of "'I-Lys- plasminogen or '2SI-plasmin with THP-1 cells, several degradative fragments were apparent in their condi- tioned media. The smaller degradative fragments (28 and 36 kDa) lacked cell binding activity and were dem- onstrated to be active by casein-zymography. A 48-kDa fragment bound to cells in a lysine-dependent manner but was not active. In contrast, phenylmethylsulfonyl fluoride-inactivated '251-plasmin retained its binding ac- tivity over 24 h, and degradative fragments were not present in the conditioned media. The binding of lZ5I- Lys-plasmin(ogen) to THP-1 cells was also examined in the presence of excess a2 plasmin inhibitor. Despite the absence of fluid-phase plasmin activity, membrane- bound '261-Lys-plasmin(ogen) decreased over 24 h. At 24 h a radiolabeled 48-kDa fragment was observed in the conditioned media and together with '251-Lys-plasmin- (ogen) was bound to cells. Unlike l2'I-Lys-p1asmin, the 48-kDa fragment did not form a complex with a2 plasmin inhibitor. Thus, autoproteolysis of receptor-bound plas- min results in fragments with truncated physiologic properties that possess either cell binding or catalytic activities. We propose that autoproteolysis is a mecha- nism for regulating membrane-bound plasmin activity.

The activation of plasminogen is an important regulatory step in cellular movement and connective tissue remodeling associated with a variety of normal and pathologic processes

* This work was supported by Research Grants R01-HL40819, PO1 HL46403, and R29 HL42606 from the National Heart, Lung and Blood Institute and Grant 4162 from the Austrian Nationalbank. 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 with 18 U.S.C. Section 1734 solely to indicate this fact.

1 To whom correspondence should be addressed: Dept. of Pathology, Cornell Medical College, 1300 York Ave., New York, NY 10021. Tel.: 212-746-6457; Fax: 212-746-8789.

(1-7). Plasmin can directly cleave the protein core of proteogly- cans (81, laminin (91, and fibronectin (9, 10). In addition, plas- min can activate a family of neutral metalloproteinases, which degrade collagens and other matrix components (11). In addi- tion to regulating cell movement and matrix remodeling, an important biological consequence of plasminogen activation is the release of matrix-bound growth factors. For example, expo- sure of endothelial cell monolayers to plasmin resulted in the release of basic fibroblast growth factor complexed to heparan sulfate proteoglycan (12). Treatment of fibroblast and fibrosar- coma cell-derived matrices with thrombin or plasmin led to the release of TGF-P1 (13): We have recently demonstrated that urokinase (uPA)-dependent plasminogen activation regulates macrophage's ability to liberate basic fibroblast growth factor and TGF-P from endothelial cell matrices (14, 15).

Macrophage plasminogen activation is localized to the peri- cellular environment by the expression of a high affinity recep- tor for uPA and relatively low affinity receptor for plasminogen (16-18). Plasminogen binding to its cellular receptor is inhib- ited by the lysine analog eaminocaproic acid (E-ACA) (16, 191, and appears to be mediated by kringles 1-3 and kringle 5 (20, 21). In contrast to the uPA receptor, at least thee candidate plasminogen receptors have been described: annexin I1 (22), a-enolase (23), and gangliosides (24). When plasminogen and uPA are bound to their respective receptors, the eficiency of plasminogen activation is increased (16-19). Following activa- tion, membrane-bound plasmin is not internalized or readily inactivated by its physiologic inhibitors (16, 17, 25, 26). Con- sequently, it would appear that cells do not possess a mecha- nism to regulate membrane-bound plasmin activity following its activation on the cell surface.

In experiments reported here, we have examined the regu- lation of membrane-bound plasmin(ogen) in human THP-1 macrophage, Results demonstrate that plasmin autoproteoly- sis and 6-ACA-resistant binding are novel regulatory mecha- nisms in the control of membrane-bound plasmin activity.

MATERIALS AND METHODS Cell Culture-Human monocyte-like THP-1 cells (27) were obtained

from American Type Tissue Culture (Rockville, MD). THP-1 cells were maintained in suspension culture in Roswell Park memorial medium (RPMI; without HEPES) supplemented with 10% Cellect Gold fetal bovine serum (FBS), penicillin (100 unitdml), streptomycin (100 pg/ml), and glutamine (ICNiFlow Laboratories, McLean, VA). During routine culture a small percentage (- 10%) of THP-1 cells spontaneously differ- entiate into an adherent cell population. The adherent THP-1 cells resemble macrophage, continue to divide, and maintain their adherent phenotype following mechanical harvesting and replating. In other

The abbreviations used are: TGF-B, transforming growth factor beta; uPA, urokinase-dependent plasminogen activation; E-ACA, €-ami-

buffered saline; MSFM, macrophage serum-free media; PMSF, phenyl- nocaproic acid; FBS, fetal bovine serum; DPBS, Dulbecco's phosphate-

methylsulfonyl fluoride; PAGE, polyacrylamide gel electrophoresis.

32660

Regulation of Macrophage Receptor-bound Plasmin 32661

0 .05 . I O .I5 .2 Bound (pMoles)

I

0 10 20 30 40 50 0.0 0.5 1.0 1.5 2.0

Plasmin (nM) Hours

FIG. 1. Dose- and time-dependent binding of plasmin to adherent THP-1 cells. Cells were plated into 96-well plates (1 x 105/well) in RPMI 1640 supplemented with 10% FBS. The next day media were removed and cells washed three times with DPBS. Left panel, cells were incubated with plasmin (0.14.0 pg/ml) at 37 "C for 2 h (m) or plasmin and 25 m E-ACA to block lysine-dependent binding (0). Right panel, cells were incubated with 1.0 pg/ml plasmin for 0-2 h (m) followed by a 10-min exposure to E-ACAto displace membrane-bound plasmin (0). Membrane-bound plasmin was determined as described under "Materials and Methods." Data represent the mean -c S.E. of six separate wells.

studies, we have measured THP-1 expression of uPA and the uPA re- ceptor.2 All of the plasminogen activator activity expressed by THP-1 cells is blocked by anti-uPA IgG?

Quantification of Plasmin Actiuity-Plasmin activity was quanti- tated by measuring the increase in fluorescence following the cleavage of the plasmin substrate u-Val-Leu-Lys-amino methyl coumarin (En- zyme Systems Products, Dublin, CA) as described previously (15). Ali- quots of conditioned media were added to microtiter wells containing 82 pl of DPBS, 0.05% Tween-20 containing 13 pg of the plasmin substrate. Samples were mixed and incubated at 37 "C for 2.5 h. Cleavage of the substrate was monitored in a Fluoroscan microplate reader (excitation, 330-380 nm; emission, 430-530 nm). Plasmin activity in the test samples was extrapolated from a standard curve utilizing 0-40 ng/ml human plasmin (American Diagnostica, Greenwich, CT).

Membrane-bound plasmin activity was quantitated by incubating THP-1 cells with either human plasmin or plasminogen in macrophage serum-free media (MSFM, Life Technologies, Inc.) for 2 h at 37 "C. Unbound plasmin was removed, and cells were washed with DPBS. MSFM containing u-Val-Leu-Lys-amino methyl coumarin was added and allowed to incubate 2.5 h. Fluorescence was monitored in a Fluo- roscan microplate reader and extrapolated to the fluorescence gener- ated by 0-40 ng/ml plasmin prepared in MSFM. The catalytic efficiency of membrane-bound plasmin has been reported to be increased relative to fluid-phase plasmin (29). We have not examined this possibility di- rectly since we were interested in relative changes in membrane-bound plasmin over time. Nonetheless, the amounts of membrane-bound plas- min may be overestimated since it was calculated by extrapolation from a standard curve of fluid-phase plasmin. It is important to note, how- ever, that the binding parameters generated via extrapolation of mem- brane-bound plasmin activity to plasmin activity in the fluid-phase are similar to that observed with '251-Lys-plasminogen (see "Results").

Iodination of Lys-plasminogen and Plasmin-Human Lys-plasmino- gen (Immuno A.G., Vienna, Austria) and plasmin (American Diagnos- tics) were iodinated according to the method of McFarlane (30). Iodi- nated proteins were separated from unincorporated '''I by gel chromatography on a PD-10 column (Pharmacia Biotech Inc.) pre- equilibrated with HEPES-buffered saline (137 mM NaCl, 4 mM KCl, 11 m glucose, and 11 mM HEPES, pH 7.4) containing 0.5% human albu- min. '251-Plasmin was irreversibly inactivated by treatment with phen- ylmethylsulfonyl fluoride (PMSF). PMSF inactivation of '251-plasmin was >90% as determined by cleavage of the plasmin substrate.

Analysis of Plasmin Fragments by SDS-PAGE and Autoradiogra- phy-THP-1 cells were incubated with '251-Lys-plasminogen or '251-plas- min in MSFM for 1 and 24 h. Aliquots of conditioned media were diluted 1:4 with SDS-sample buffer without P-mercaptoethanol and boiled 3 min. Samples and molecular weight markers were applied to 415% gradient gels. Gels were fured in 10% acetic acid for 1 h and stained with Coomassie Blue overnight. Following destaining the gels were dried and placed on Kodak XOMAT AR film with a Cronex (DuPont) enhanc- ing screen.

Enzyme Zymography-Conditioned media were concentrated in an ice bath utilizing an Amicon ultrafiltration chamber with a Y"10 (10

D. J. Falcone, T. A. McCaffrey, W. Borth, J. Mathew, and K. McAdam, manuscript submitted for publication.

kDa cutoff membrane). SDS sample buffer without mercaptoethanol was added to the media samples and heated for 30 min at 37 "C. Samples and molecular weight markers were electrophoresed in a 12% polyacrylamide gel containing 0.1% casein. The gel was then washed (two times) in 2.5% Triton X-100 to remove SDS. The gel was incubated at 37 "C for 40 h in 50 mM Tris-HC1, 10 mM CaCl,, pH 7.8, and stained with Coomassie Blue. The presence of caseinolytic activity was identi- fied as clear bands on a uniform blue background following destaining.

Western Blot for Plasmin-Cell monolayers were lysed with SDS sample buffer without mercaptoethanol and immersed in boiling water for 5 min. Biotinylated molecular weight markers and samples were electrophoresed in P15% polyacrylamide gradient gels. Proteins were transferred to a nitrocellulose membrane, following which the mem- brane was blocked in Tris-buffered saline, 0.05% Tween-20 (TTBS) con- taining 5% dry defatted milk for 1 h. Following two washes ("BS), the membrane was incubated with blocking buffer containing 10 pg/ml monoclonal antibody directed against kringles 1-3 of human plasmin IgG for 1 h (American Diagnostica). The membrane was washed (two times; TTBS) and incubated for 1 h with 10 pg/ml biotinylated rabbit anti-mouse IgG (Pierce) in blocking buffer. The membrane was then washed (two times; TTBS) and incubated 1 h with 4.6 pg/ml avidin- horseradish peroxidase (Pierce) in TBS, 0.1% gelatin. Diaminobenzi- dine was used as a substrate.

RESULTS

Binding of Plasmin and Plasminogen to THP-1 Cells-A va- riety of cell types, including monocytes, express large numbers of relatively low affinity binding sites for plasminogen (16-24, 26). The binding of plasminogen appears to be lysine dependent since the lysine analog E-ACA competitively blocks plasmino- gen binding to cells. In these experiments, we have measured the binding of catalytically active plasmin to adherent THP-1 macrophage. As seen in Fig. 1, plasmin binds to adherent THP-1 cells in a dose- and time-dependent manner. The dose- dependent plasmin binding is completely blocked if E-ACA is present during the incubation period. Based on Scatchard anal- ysis of four separate experiments, the estimated Kd of binding was determined to be 15.8 2 6.2 r m (mean 2 S.E.) with 1.4 2 0.26 x lo6 binding siteskell. The observed binding affinity of plasmin to the THP-1 macrophage was significantly greater than reported for plasminogen binding to either endothelial cells (Kd = 310 2 235 m) (19) or suspension cultures of U937 monocytes (Kd = 800 * 500 nM) (16). However, it was of similar magnitude to that reported for plasmin binding to endothelial cells (Kd = 77 2 18 nM) (19). The plasminogen in the FBS used to grow the THP-1 cells did not affect plasmin(ogen) binding. Results of at least 20 experiments have demonstrated that THP-1 cells incubated with 10% FBS alone had trace or no detectable plasmin(ogen) on their surfaces by Western blot and/or activity measurements (data not shown).

Although E-ACA competitively inhibits plasmin binding to

32662 Regulation of Macrophage Receptor-bound Plasmin

0 Total Bound

1 4 8 24

Hours

THP-1 cells. Cells were plated into 96-well plates (1 x 105/well) in FIG. 2. Loss of c-ACA-sensitive plasmin binding to adherent

RPMI 1640 supplemented with 10% FBS. The next day media were removed and cells washed with DPBS. Media was replaced with MSFM containing 1 pg/ml plasmin and incubated 1-24 h followed by a 10-min

bound plasmin was determined as described under "Materials and exposure to E-ACA to displace membrane-bound plasmin. Membrane-

Methods." Data represent the mean * S.E. of six separate wells.

adherent THP-1 cells (Fig. 1, left panel), the fraction of mem- brane-bound plasmin that is resistant to subsequent displace- ment with E-ACA increases with time (Fig. 1, right panel). In addition, adherent THP-1 cells were incubated with plasmin for 1-24 h and membrane-bound plasmin determined. There was a time-dependent loss of THP-1 membrane-bound plasmin activity (Fig. 2). Following 24 h of incubation, -80% of the membrane-bound plasmin activity observed at 1 h was lost. Furthermore, the remaining membrane-bound plasmin (20%) was not susceptible to displacement by E-ACA.

In an effort to identify the mechanism(s) responsible for the observed loss of enzymatically active plasmin from the THP-1 surface, we determined whether the binding properties of the THP-1 plasmin(ogen) receptor were sensitive to plasmin cleav- age. For this purpose, the binding of 1251-Lys-plasminogen to cells preincubated with plasmin was determined (Fig. 3). Lys- plasminogen is the plasmin-modified form of Glu-plasminogen (26, 31). Lys-plasminogen is formed at the cell surface and is the form of plasminogen that is preferentially bound (26). The binding of '251-Lys-plasminogen to plasmin-treated THP-1 cells was similar to the binding of plasmin to untreated cells (Fig. 1). Based on Scatchard analysis of four separate experi- ments, the estimated Kd of binding is 14.7 5 1.41 m, and the numbers of binding sites is 1.96 5 0.30 x 106/cell. Therefore, plasmin-mediated proteolysis of the plasminogen receptor could not be responsible for the observed loss of plasmin from the cell surface.

Autoproteolysis Leads to Catalytically Active Plasmin with- out Cell Binding Activity-We determined whether the region of plasmin responsible for cell binding is lost or modified during an extended incubation with adherent THP-1 macrophage. Cells were incubated with plasmin (1 pgfml) for 24 h after which their conditioned media were recovered and assayed for remaining plasmin activity. The binding of 24 h conditioned plasmin to freshly plated THP-1 cells was compared to control plasmin (Table I). THP-1 cells incubated with control plasmin for 1 h bound >lOO-fold more plasmin than cells incubated with conditioned plasmin. These data suggest that either cells are releasing an inhibitor of plasmin binding, or plasmin itself is modified during an extended incubation with the THP-1 cells. In order to determine whether THP-1 cells released an inhibi- tor of plasmin binding to its receptor, fresh plasmin was added to the 24-h conditioned media and membrane-bound plasmin determined. As seen in Table I, the binding of plasmin to THP-1 cells in the presence of conditioned media was restored by the

25 - 1 - - 3 2 0 -

b z m 15 - - -0 c 10 -

K t 5 - 0 .I .2 .3 .4

Bound (pMole8)

0 15 30 45 BO 76

126 I-Lys-Plg (nM)

FIG. 3. The effect of plasmin pretreatment on binding of '%SI-

Lys-plasminogen to adherent THP-1 cells. Cells were plated into 96-well plates (1 x 105/well) in RPMI 1640 supplemented with 10% FBS. Following adherence, the media were removed and replaced with

incubated with 0.1-6 pdml '261-Lys-plasminogen for 1 h at 4 "C. Fol- MSFM containing 1 pgfml plasmin. The next day cells were washed and

lowing incubation with labeled plasminogen, cells were washed and membrane-bound plasminogen was released by exposure to 25 mM E-ACA (10 min). 1251-Lys-plasminogen resistant to a-ACA displacement was recovered by treating the monolayer with 0.5 NaOH. Data repre- sent the mean -c S.E. of three separate samples.

TABLE I Loss of plasmin's cell binding activity

(105/well) in MSFM. Cells were incubated with plasmin for 1 or 24 h Plasmin (1 pgfml) was added to adherent cultures of THP-1 cells

following which the media were recovered and assayed for plasmin activity. The cells were washed; membrane-bound plasmin activity was quantitated as described under "Materials and Methods." To determine whether the 24-h conditioned media contained an inhibitor of plasmin binding, fresh plasmin (Pls to) was added and incubated with THP-1 cells for 1 h. Data represent the mean * S.E. of six wells.

Media activity p g f ml Pls t, 1000 Pls tZ4 946 f 75

Membrane-bound activity ngl ml THP-1 0.03 * 0.1 THP-1 + Pls t, 16.69 * 0.05 THP-1 + Pls h4 0.13 * 0.01

THP-1 + Pls to 13.37 * 0.54 THP-1 + Pls h4

addition of fresh plasmin. Furthermore, the ability of plasmin to bind to cells was lost when plasmin was incubated in the absence of cells for 24 h. Therefore, we conclude that plasmin can form an enzymatically active species that is not recognized by the plasmin(ogen) receptors.

A possible mechanism for the observed loss in plasmin's cell binding properties is autoproteolysis. Therefore, we examined by SDS-PAGE-autoradiography alterations in '261-Lys-plasmin- ogen incubated with adherent THP-1 cells for 1 and 24 h. Un- der nonreducing conditions, the major fraction of labeled plas- minogen in the 1-h conditioned media co-migrated with intact '251-Lys-plasminogen (Fig. 4). Radiolabeled degradative frag- ments between 48 and 40 kDa, which were absent in the start- ing material, became visible following 1 h of incubation with THP-1 cells. In contrast, at 24 h the majority of the radiola- beled plasminogen had shifted to the 48- and 40-kDa peptide fragments. Smaller degradative fragments, as well as intact '251-Lys-plasmin(ogen) were faintly visible at 24 h.

Identification of Active Plasmin Fragments by Zymo- graphy-In order to determine which of the plasmin fragments in THP-1-conditioned media were enzymatically active, cells were incubated with unlabeled plasminogen for 24 h and ana- lyzed by SDS-PAGE zymography. Plasmin fragments express- ing proteolytic activity degrade casein included in the

Regulation of Macrophage Receptor-bound Plasm.in

Hours 1 24 "

kD

-205 -1 17

'251-Lys-Plg -80 -50

-33 -28 -19

FIG. 4. Degradative alterations of "511-Lys-planminopn incu- bated with THP-1 cells. Cells wrre platrd into 12-well plates ( 5 x I0"iwell) as drscrihed in Fig. 1. Following adhcrrnce the crlls were washrd and mrdia replaced with MSFM containing 1 pg/ml 12"I-Lys- plasminogen. Conditioned medium was recoverrd at 1 and 24 h. Alter- ations in molecular weight of "~'1-Lys-Plasminogen was determined by SDS-PACE followed by autoradiography.

polyacrylamide gel. Areas of casein digestion appear unstained following staining of the gel with Coomassie Blue. As positive controls, both plasminogen and plasminogen preincubated with uPA for 1 h were examined. Plasmin activity in the plasmino- gen preparation was clearly evident in the zymograph (Fig. 5). When plasminogen was activated by 1 h of preincubation with uPA, a wide range (80-20 kDa) of enzymatically active plasmin fragments were formed. In contrast, conditioned media recov- ered from THP-1 cells incubated with plasminogen for 24 h contained fewer degradative forms ofenzymatically active plas- min. Several high and two lower molecular weight caseinolytic bands were observed. The lower molecular mass caseinolytic bands (28 and 36 kDa) correspond to the low molecular mass 12'I-labeled plasminogen fragments observed in the media of THP-1 cells a t 24 h (Pig. 4). However, caseinolytic activity did not co-migrate with a prominent radiolabeled fragment of plas- minogen tie. 48 kDa). When media derived from THP-1 cells cultured in the absence of plasminogen were analyzed by SDS- PAGE zymography, no caseinolytic bands were observed (data not shown). Taken together, (Table I, Figs. 4 and 5), these data demonstrate that following activation, plasmin autoproteolysis results in the generation of proteolytically active fragments that do not express cell binding properties.

PMSF-inactivated '""I-Plasmin Does Not Lose Its Cell Rind- ing Properties-If autoproteolysis is responsible for the loss in cell-bound plasmin, then the binding of inactivated l2'II-plasmin to cells would not be expected to change over time. Therefore, we compared the binding of "%plasmin and PMSF-inactivated "'I-plasmin to adherent THP-1 cells a t 1 and 24 h (Fig. 6). By autoradiography, the radiolabeled plasmin preparation con- tained a 48-kDa degradative fragment. A similar degradative fragment was observed in the conditioned media when THP-1 cells were incubated with l2'1-Lys-plasminogen for 24 h (Fig. 4). Following a 1-h incubation, both intact plasmin and the 48-kDa degradative fragment were found associated with the adherent THP-1 cells. Exposure of the cells with c-ACA released a ma- jority of the intact plasmin and the 48-kDa radiolabeled frag- ment. As observed utilizing '251-Lys-plasminogen as a tracer

205 - 117- 80 -

50 -

32663

kD

-106 -80

-50

-33

-28

FIG. 5. Zymographir idcmtifie;ation of c.nzyrn:ttic.;ally active plasmin frrcgmrnts. ('rlls w n . plated into 'r-25 flasks 1: x 10") as descrihrd in FIG. 1. Following ndhcwnce the cc~lls w r c - washed and media replnccd with M'SFlI containing 1 &ml pl:lsminogrn. The con- ditioned mrdium was collrcted 24 h later and concrntratcd 5:l by ul- trafiltration as described undrr "Materials and Methods." SDS sample huffrr was added to aliquots of TIIP-1-conditionrtl mrdium, plasmino- gen and plasminogen prrincuhatrd (1 h, with high molecular Wright UPA. Samplrs were electrophoresed in composite polyacrylnmide-casc,in gels. Activr plasmin fragmrnls wrre identified by zymomaphy.

Lysate Media 1-" 24 1 4 (hrs)

kD -200

-46 -30

- + - + E-ACA

Lvsate Media

j-200

Inactive '251-Pls

1-97

4-69 ~ -46

-30

- + - + E-ACA FIG. 6. Effect of PMSF inactivation on the binding of '061-plas-

min to THP-I cells. (klls wrrr platrd into 6-wll platrs (2 x 10'iiwcll) as descrihcd in Fig. 1. Following adhrrence the cells were washed and media rrplaced with MSFM containing 1 pg/ml ""I-plasmin or PMSF- inactivated "'T-plasmin. Conditionrd media and cell lysates were re- covered at 1 and 24 h. Alterations in molrcular weight of "'I-plasmin was determined by SDS-PACE followed hy autoradiomaphy.

(Fig. 4), a t 24 h the majority of the 12'II-plasmin was recovered in the 48-, 40-kDa, and several other smaller degradative frag- ments (Fig. 6). Associated with the formation of degradative fragments of '2'I-plasmin, a t 24 h '2'I-plasmin recovered from THP-1 cells was dramatically reduced as compared to 1 h. In

32664 Regulation of Macrophage Receptor-bound Plasmin

kD

- 200

- 97

- 66

- 45

24 6 4 2 1 PIS Hours

FIG. 7. The loss of anti-kringle IgC binding to plasmin. Cells were platrd into 12-well plates (0.5 x IOfl/'/well~ as descrihrd in Fig. 1. Following adherence the crlls were washed and media replaced with MSFM containing 1 pg/ml plasmin. Conditioned media wrrr rrcoverrd at 1-24 h and the immunoreactivity of recovered plasmin was deter- mined by Wrstrrn blot utilizing monoclonal anti-kringlr I&.

marked contrast, no degradative change in the active site in- hibited '251-plasmin was observed in the media over 24 h, and similar amounts of PMSF-inactivated 125"Iplasmin were bound to cells following 1 and 24 h of incubation (Fig. 6). The e-ACA- resistant fraction of PMSF-plasmin was comparable to that of untreated plasmin. These data suggest that autoproteolysis can regulate cell binding.

Loss of Anti-kr ingl~ IgG Binding to Autoproteolyzed Plasmin-Figs. 4 and 6 demonstrate the generation of plasmin degradation fragments over a 24-h incubation with THP-1 cells. However, in both experiments small quantities of intact plasmin(ogen) were present in the conditioned media at 24 h. The lack of substantial membrane-bound activity at 24 h (Fig. 6), despite the presence of material that co-migrates with intact plasmin(ogen), suggests that other alterations in plasmin- (ogen) may have taken place. Since the kringle domains of plasminogen's A-chain mediate its binding to both fibrin and endothelial cells (20, 21), we determined whether plasmin au- toproteolysis was associated with a loss of reactivity to a mono- clonal anti-kringle I g G . When conditioned media from THP-1 cells incubated with plasmin for 1-24 h were examined by Western blot, a clear time-dependent reduction in kringle im- munoreactivity was observed (Fig. 7). Little or no intact immu- noreactive plasmin could be demonstrated at 24 h. These data suggest that the failure of the remaining intact plasmin(ogen) in the 24-h conditioned media to bind THP-1 cells was due to alterations in its cell-binding dornainb).

Plasmin Autoproteolysis Occurs on the Cell Surface in the Presence of Excess a2 Anti-plasmin-Under physiological con- ditions fluid-phase plasmin will be rapidly inactivated by plas- min inhibitors (32). Therefore, i t is not clear from experiment.. described above whether autoproteolysis could contribute to the physiologic regulation of membrane-bound plasmin activ- ity. Consequently, we determined whether autoproteolysis would occur on the cell surface in the presence of 20-fold molar excess of a2 plasmin inhibitor. THP-1 cells were preincubated with TGF-p1 overnight to up-regulate their expression of both uPA and the uPA receptor.2 However, TGF-p1 does not affect plasminogen binding.2 When TGF-pl-primed macrophages were incubated with media containing 12511-Lys-plasminogen and a2 plasmin inhibitor for 1 h, the conditioned media con- tained two major radiolabeled bands (-80 and 140 kDa). The lower molecular weight band comigrated with intact '"'I-Lys- plasminogen. We have concluded that the higher molecular weight band is an SDS-resistant complex of "'I-Lys-plasmin and a2 plasmin inhibitor (6,570 kDa) for three reasons: a2 plasmin inhibitor rapidly forms a denaturation-resistant com- plex with plasmin (32). This band was not observed in the absence of a 2 plasmin inhibitor. Plasmin activity was undetect-

kD

204 - 132-

65 - 43 - 30 -

E-ACA

Lysates Media 1 24 1 24

~~

- + - + FIG. 8. Effect of a2 plasmin inhibitor on the binding of '=!-Lys-

plasminogen to THP-1 cells. Crlls wrre incuhatrd with TGF-fil I5 nglml) ovrrnight, harvrstrd, and platrd into 12-well plates ( 5 x 105/iwell I as descrihrd in Fig. 1. Following adherence the crlls wrre washed and media replaced with MSFM contnining TGF-131, 1 pg/ml ""I-l,vs-plas- minogen, and 20 pg/ml rr2 plasmin inhihitor. Conditionrd media were recovered a t 1 and 24 h. Alterations in molrcular weight of

autoradiography. I I' c I-Lys-Plasminogen wrrr drtrrminrd by SDS-PAGE followed hy

Fluid Phose "XF-

4 AP

-c - c 4 K K

C

I' I K r

-c Plg 2 Kringledomain

C CotaVtic domain . .-Ap A,aha p,ovnin

Inhmb.*o'

FIG. 9. A model depicting autolytic regulation of receptor- bound plasmin activity. I'lasminogrn and uPA hind to their rrspec- tivr rrcrptors. l'lasminoyrn initially hinds to tho cell surfacr in a Iysine- drpendrnt mannrr via its kringles. I'lasminogrn is preferentially cleaved hy uPA on the cell surface genrrating catalytically active plas- min. Plasmin which has dissociatrd from its rrcrptor is inhihited by (12 plasmin inhihitor. Membrane-hound plasmin is autoprotrolyzed grner- ating inactive fragments that rrmain hound to the recrptor. Catalyti- cally active fragmrnts released as a consrquencr of autoproteolysis are inhibited by a2 plasmin inhihitor. A fraction of memhranr-hound plas- min hecomes resistant to displacement hy the lysine analog cACA and is protrctrd from plasmin drgradation.

able in the media samples containing a2 plasmin inhibitor (data not shown). At 1 h intact "'I-Lys-plasmin(ogen) was found associated with the THP-1 cells. Exposure of the cells with c-ACA released a majority of the bound I:!"I-Lys-plasmin- ogen, however, these differences are not readily apparent be- cause films were overexposed to visualize other radioactive fragments.

The 24-h conditioned media also contained intact '"I-Lys- plasminogen and complexes of 1z51-Lys-plasmin and a2 plasmin inhibitor; however, the proportion of '2511-Lys-plasmin associ- ated with a2 plasmin inhibitor was increased. In addition, de- spite the absence of detectable plasmin activity in the fluid- phase, degradative fragments of ""I-Lys-plasminogen were

Regulation of Macrophage Receptor-bound Plasmin 32665

visible. Similar degradation fragments of greater intensity were observed in the absence of a2 plasmin inhibitor. The amount of '251-Lys-plasminogen bound to cells at 24 h was markedly reduced. In addition to intact '251-Lys-plasmin(ogen), a radiolabeled fragment also observed in the media was bound to the cells. Both the intact 1251-Lys-plasminogen and the frag- ment were displaced by exposure to E-ACA indicating lysine- dependent binding. Since fluid-phase plasmin activity was completely inhibited by the excess of a2 plasmin inhibitor, these data demonstrate that plasmin autoproteolysis occurs on the cell surface despite the presence of excess plasmin inhibitor.

DISCUSSION

The broad substrate specificities of plasmin require that cells localize its activation in order to prevent collateral tissue dam- age. The directed activation of plasminogen by macrophage is affected by their expression of receptors for uPA and plasmin- ogen (6, 16-18, 33). The regulation of these receptors and the fate of ligands bound to them have great implications for a diverse group of macrophage functions including matrix remod- eling, migration, and mobilization of matrix-bound growth factors.

In contrast to the uPA receptor (34-37), little has been re- ported concerning regulation of plasminogen receptors (38). In these studies we demonstrate that plasmin is subject to auto- proteolysis which leads to a loss in receptor recognition but not catalytic activity. When THP-1 macrophage were incubated with either plasmin or plasminogen in serum-free media, mem- brane-bound plasmin activity declined dramatically over 24 h despite the presence of catalytically active plasmin in the me- dia (Fig. 2, Table I). Plasmin-mediated proteolysis of the plas- midogen) receptor was not responsible for the loss in mem- brane-bound plasmin activity since preincubation of THP-1 cells with plasmin did not affect binding of 1251-Lys-plasmino- gen (Fig. 3). When conditioned media from cells incubated with either '251-Lys-plasminogen or 1251-plasmin were examined by SDS-PAGE and autoradiography, autoproteolysis of the radio- labeled material had clearly taken place (Figs. 4 and 6). Fol- lowing a 24-h incubation with THP-1 cells, two major degrada- tive fragments of 'z51-Lys plasmin(ogen1 were visible at 48 and 40 kDa. In addition, faintly visible intact '251-plasmin(ogen) and small degradative fragments were observed. Similar frag- ments were observed when cells were incubated with '251-plas- min. Both intact '251-plasmin(ogen) and the 48-kDa fragment were found associated with the THP-1 cells and could be dis- placed by e-ACA (Fig. 6). Although the 48-kDa plasmin frag- ment bound THP-1 cells in a lysine-dependent manner, it did not appear to possess plasmin activity as judged by casein zymography (Fig. 5) . In contrast to either '251-Lys-plasminogen and 1251-plasmin, degradative fragments did not form when cells were incubated with PMSF-inactivated 1251-plasmin. Fur- thermore, similar amounts of PMSF-inactivated '251-plasmin were bound to cells at 1 and 24 h (Fig. 6). Thus, it appears that under serum-free conditions, autoproteolysis dissociates plas- min into fragments with truncated physiologic properties that possess either cell binding or catalytic activities.

The modification of structure and biological properties of native (Glu'l-plasminogen (92 kDa) by plasmin and other pro- teases has been described previously. A "preactivated form of plasminogen termed Ly~~~-plasminogen (84 kDa) is obtained by plasmin-mediated cleavage of the NH,-terminal 76 amino acid peptide (39). Ly~~~-plasminogen is formed on the cell surface and binds to cells with higher affinity than G1u'-plasminogen (26, 31). Likewise, the kinetics of Lys-plasminogen conversion to Ly~~~-p la smin by tissue plasminogen activator is markedly enhanced (26). When plasmin was incubated in a protein-free aqueous buffer (pH 6.5), its catalytic activity decreases rapidly

over time (40). The observed decrease in catalytic activity was due to autodegradation of the B-chain (26.5 ma) which con- tains the catalytic domain of plasmin(ogen1 (40). In contrast, most of the A-chain (63 m a ) , which contains plasmin's five kringle domains, remained unaltered (40). Micro-plasmin (in- tact B-chain; 26.5 kDa) was obtained by incubating plasmin in a protein-free aqueous buffer at pH 11 (39,401. At an alkaline pH, the B-chain was protected from degradation, whereas the A-chain was partially degraded to a protein of 58 kDa (40,411. This partially degraded A-chain bound lysine-Sepharose dem- onstrating the preservation of functional kringle(s) that medi- ate lysine-dependent binding (40). Mini-plasminogen (38 kDa) containing kringle 5 and the B-chain was obtained by digestion of plasminogen with elastase (42).

Lysine-dependent plasminogen binding to both fibrin and endothelial cells is inhibited by kringles 1-5 (55 kDa) and mini- plasminogen, but not micro-plasminogen or kringles 1-3 (20, 21,43). Kringle 4 does not exhibit significant binding to either fibrin or endothelial cells (20,43). These data demonstrate that kringle 5 plays a pivotal role in the lysine-dependent binding of plasmin(ogen1 to the endothelial cell surface and fibrin. In ex- periments reported here, autoproteolysis results in an enzy- matically inactive plasmin fragment (48 kDa) that binds cells in a lysine-dependent manner, and catalytically active plasmin fragments (-28 and 36 kDa) which are devoid of cell binding properties. These active fragments may represent micro-plas- min or mini-plasmin with an altered kringle conformation.

Plasminogen receptors provide a mechanism for localization of zymogen, its activation, and protection from inactivation (16, 19, 25). Since fluid-phase plasmin is rapidly inhibited by a2 plasmin inhibitor (k,,, 2.8 x M -' s -' (321, it is not clear if autoproteolysis could play a role in regulating membrane- bound plasmin activity under physiologic conditions. There- fore, we examined whether autoproteolysis could take place in the presence of excess a2 plasmin inhibitor. When THP-1 cells were incubated with '251-Lys-plasminogen and the plasmin in- hibitor, intact membrane-bound '251-plasmin(ogen) decreased dramatically over 24 h despite the absence of detectable fluid- phase plasmin activity (Fig. 8). As observed in the absence of the plasmin inhibitor, intact '251-Lys-plasmin(ogen) and a ra- diolabeled 48-kDa fragment observed in the conditioned media were bound to the cells. The binding of both '251-Lys-plasmin- (ogen) and the 48-kDa fragment were lysine dependent since they were displaced by E-ACA. The failure of the fragment to form a complex with a2 plasmin inhibitor corroborates our conclusion that it lacks an active site. Smaller radiolabeled fragments which were observed in the absence of a2 plasmin inhibitor and demonstrated to be active by zymography were not present. These data demonstrate that in the presence of plasmin inhibitors, plasmin autoproteolysis is restricted to the cell surface and generates fragments that possess either cell binding or catalytic activities.

In these studies, we have also demonstrated that plasmin binding to THP-1 cells is initially E-ACA sensitive but becomes resistant to displacement with E-ACA over time. Following a 24-h incubation with plasmin or '251-Lys-plasminogen, all mem- brane-bound plasmin activity or '251-Lys-plasmin(ogen) was re- sistant to E-ACA displacement and further autoproteolysis. I t has been recently reported that endothelial cells in suspension cross-linked plasmin to their surfaces via expression of trans- glutaminase activity (28). However, in these studies the inabil- ity of E-ACA to displace plasmin bound to adherent THP-1 was not due to transglutaminase-mediated cross-linking, since cell- associated large molecular weight complexes containing lZ5I- Lys-plasmin(ogen) were not observed in SDS-PAGE. Thus, E-ACA-resistant and transglutaminase-independent binding of

32666 Regulation of Macrophage

plasmin represents an undefined mechanism for localizing plasmin to the cell surface.

The current model of of plasminogen binding and activation provides mechanisms for localization of zymogen, its regulated activation, and protection from inactivation. However, the model does not provide a mechanism for cell clearance of re- ceptor-bound protease. Based on results of experiments re- ported here, we propose that autoproteolysis is a mechanism for regulating receptor-bound plasmin activity under condi- tions of high pericellular plasmin concentration (Fig. 9).

REFERENCES 1. Dent, M. A. R., Sumi, R., Morris, J., and Seeley, P. J. (1993) Eur. J. Neurosci.

2. Konno, H., Tanaka, T., Maruo, Y., Nishino, N., Nakamura, S., Baba, S., and

3. Huarte, J., Vassalli, J.-D., Belin, D., and Sakkas, D. (1993) Deu. Biol. 157,

4. Mohanam, S., Sawaya, R., McCutcheon, I., Mi-Osman, Boyd, D., and Rao, J. S.

6. Estreicher, A,, Muhlhauser, J., Carpentier, J. L., Orci, L., and Vassalli, J. D. 5. McNeill, H., and Jensen, P. J. (1990) Cell Regul. 1, 843-852

7. Schlecte, W., and Brattain, M., and Boyd, D. (1990) Cancer Commun. 2, 173-

8. Richardson, M., Hatton, M. W., and Moore, S. (1988) Clin. Znuest. Med. 11,

9. Moser, T. L., Enghild, J. J., Pizzo, S. V., and Stack, M. S. (1993) J. Biol. Chem.

10. Wang, L.. Lin, X., Cui, D., Wang, Z., and Chi, C. (1994) J. Biol. Chem. 269,

11. Murphy, G., Atkinson, S., Ward, R., Gavrilovic, R., and Reynolds, J. J. (1992)

5,633-647

Takada, A. (1993) Eur. Surg. Res. 25,239-244

539-546

(1993) Cancer Res. 53,4143-4147

(1990) J. Cell Biol. 111, 783-792

179

139-150

268, 18917-18923

43324336

12. Saksela, O., and Riflin, D. P. (1990) J. Cell Biol. 110, 767-775 13 Tninale. .J.. Knli. K.. and Keski-Oia. J . (1992) J. Biol. Chem. 267.25378-25384

Ann. N . Y Acad. Sci. 667,l-12

". ".

14. Falcone, D. J., McCaffrey, T. A,, HaimovitzIFriedman, A,, VerGlio, J. A,, and

15. Falcone, D. J., McCaffrey, T. A,, Haimovitz-Friedman, A,, Vergilio, J. A., and

16. Plow, E. E , Freany, D. E., Plescia, J., and Miles, L. A. (1990) J. Cell Biol. 103,

17. Kirchheimer, J. C. , and Remold, H. G. (1989) Blood 74, 1396-1402 18. Ellis, V., Behendt, N., and Dano, K (1991) J. Biol. Chem. 266, 12752-12758

= ~ ~ ~ . , . , ~ " ~ .,", "" ~~~~~~~ .

Nicholson, A. C. (1993) J. Biol. Chem. 268, 11951-11958

Garcia, M. (1993) J. Cell. Physiol. 155, 595-605

2411-2420

Receptor-bound Plasmin 19. Hajjar, K A., Harpel, P. C., Jaffe, E. A,, and Nachman, R. L. (1986) J. Biol.

Chem. 261,11656-11662 20. Wu, H. L., Wu, LS., Fang, R.-Y., Hau, J. S., Wu, D. H., Chang, B. I., Lin, T. M.,

21. Burge, S. M., Marshall, J. M., and Cederholm-Williams, S. A. (1992) BE J. and Shi, G. Y. (1992) Biochem. Biophys Res. Commun. 188, 703-711

22. Hajjar, K. A,, Jacovina,A. T., and Chacko, J. (1994)J. Biol. C h m . 269,21191- Dermatol. 126.35-41

21197 23. Miles, L. A,, Dahlberg, C. M., Plescia, J., Felez, J., Kato, K., and Plow, E. F.

24. Miles, L. A,, Dahlberg, C. M., Levin, E. G., and Plow, E. F. (1989) Biochemistry (1991) Biochemistry 30,1682-1691

25. Hall, S. W., Humphies, J. E., and Gonias, S. L. (1991) J. Biol. Chem. 266, 28,9337-9343

26. Hajjar, K. A,, and Nachman, R. L. (1988) J. Clin. Inuest. 82, 1769-1778 12329-12336

27. Tsuchiya, S., Yamabe, M., Yamaguchi, Y., Kobayashi, Y., Konno, T., and Tada,

28. Bendixen, E., Borth, W., and Harpel, P. C. (1993) J. Biol. Chem. 268, 21962-

29. Gonzalez-Gronow, M., Stack, S., and Pizzo, S. V (1991) Arch. Biochem. Bio-

30. McFarlane, A. S. (1958) Nature 162.53 31. Silverstein, R. L., Friedlander, R. J., Nicholas, R. L., and Nachman, R. L.

32. Travis, J., and Salvesen, G. S. (1983) Annu. Reu. Biochem. 52,655-709 (1988) J. Clin. Invest. 82, 194LL1955

33. Kirchheimer, J. C., and Remold, H. G. (1989) J. Immunol. 143,2634-2639 34. Ploug, M., Ronne, E., Behendt, N., Jensen, A. L., Blasi, F., and Dan0 K. (1991)

35. Lund, L. R., Ronne, E., Roldan, A. L., Behendt, N., Romer, J., Blasi, F., and J. Biol. Chem. 266, 1926-1933

36. Picone, R., Kajtaniak, E. L., Nielsen, L. S., Behendt, N., Mastronicola, M. R., Dano, K. (1991) J. Biol. Chem. 266,5177-5181

Cubellis, M. V., Stoppelli, M. P., Anderson, S., Dano, K., and Blasi, F. (1989) J. Cell Biol. 108, 693-702

37. Mignatti, P., Mazzieri, R., and Riflrin, D. B. (1991) J. Cell Biol. 113,1193-1201 38. Felez, J., Miles, L. A,, Plescia, and Plow, E. E (1990) J. Cell Biol. 111, 1673-

39. Castellino, F. J., Ploplis, V., Powell, J. R., and Strickland, D. (1981) J. Biol.

40. Wu, L.-H., Shi, G. Y., and Bender, M. L. (1987) Proc. Nutl. Acud. Sci. U. S. A.

41. Wu, L.-H., Shi, G. Y., Wohl, R. C., and Bender, M. L. (1987) Proc. Natl. Acad.

42. Sottrup-Jensen, L., Claeys, H., Zajdel, M., Petersen, T. E., and Magnusson, S.

43. Wu, H.-L., Chang, B.-I., Wu, D.-H., Chang, L.-C., Gong, C.-C., Lou, K-L., and

K. (1980) Znt. J. Cancer 26, 171-176

21967

phys. 286,625428

1683

Chem. 256,4778-4782

84,829%8295

Sci. U. S. A. 84,87934795

(1978) Prog. Chem. Fibrin. Thrombo. 3, 191-209

Shi, G.-Y. (1990) J. Biol. Chem. 265, 19658-19664