doi:10.1182/blood.V99.8.28352002 99: 2835-2844   

 Carmeliet and Pura Muñoz-CánovesMònica Suelves, Roser López-Alemany, Frederic Llui?s, Gloria Aniorte, Erika Serrano, Maribel Parra, Peter regeneration in vivoPlasmin activity is required for myogenesis in vitro and skeletal muscle

http://bloodjournal.hematologylibrary.org/content/99/8/2835.full.htmlUpdated information and services can be found at:

(2497 articles)Hemostasis, Thrombosis, and Vascular Biology   �Articles on similar topics can be found in the following Blood collections

http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#repub_requestsInformation about reproducing this article in parts or in its entirety may be found online at:

http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#reprintsInformation about ordering reprints may be found online at:

http://bloodjournal.hematologylibrary.org/site/subscriptions/index.xhtmlInformation about subscriptions and ASH membership may be found online at:

Copyright 2011 by The American Society of Hematology; all rights reserved.American Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the    

For personal use only. by guest on June 12, 2013. bloodjournal.hematologylibrary.orgFrom

HEMOSTASIS, THROMBOSIS, AND VASCULAR BIOLOGY

Plasmin activity is required for myogenesis in vitro and skeletal muscleregeneration in vivoMonica Suelves, Roser Lopez-Alemany, Frederic Lluıs, Gloria Aniorte, Erika Serrano, Maribel Parra,Peter Carmeliet, and Pura Munoz-Canoves

Plasmin, the primary fibrinolytic enzyme,has a broad substrate spectrum and isimplicated in biologic processes depen-dent upon proteolytic activity, such astissue remodeling and cell migration. Ac-tive plasmin is generated from proteolyticcleavage of the zymogen plasminogen(Plg) by urokinase-type plasminogen acti-vator (uPA) and tissue-type plasminogenactivator (tPA). Here, we have investi-gated the role of plasmin in C2C12 myo-blast fusion and differentiation in vitro, aswell as in skeletal muscle regeneration invivo, in wild-type and Plg-deficient mice.Wild-type mice completely repaired ex-perimentally damaged skeletal muscle. Incontrast, Plg 2/2 mice presented a severe

regeneration defect with decreased re-cruitment of blood-derived monocytesand lymphocytes to the site of injury andpersistent myotube degeneration. In addi-tion, Plg-deficient mice accumulated fi-brin in the degenerating muscle fibers;however, fibrinogen depletion of Plg-deficient mice resulted in a correction ofthe muscular regeneration defect. Be-cause we found that uPA, but not tPA, wasinduced in skeletal muscle regeneration,and persistent fibrin deposition was alsoreproducible in uPA-deficient mice follow-ing injury, we propose that fibrinolysis byuPA-dependent plasmin activity plays afundamental role in skeletal muscle regen-eration. In summary, we identify plasmin

as a critical component of the mammalianskeletal muscle regeneration process,possibly by preventing intramuscular fi-brin accumulation and by contributing tothe adequate inflammatory response af-ter injury. Finally, we found that inhibitionof plasmin activity with a2-antiplasminresulted in decreased myoblast fusionand differentiation in vitro. Altogether,these studies demonstrate the require-ment of plasmin during myogenesis invitro and muscle regeneration in vivo.(Blood. 2002;99:2835-2844)

© 2002 by The American Society of Hematology

Introduction

The proteolytic conversion of the ubiquitous zymogen plasmino-gen (Plg) to the active protease plasmin is an extensively usedmechanism for the generation of extracellular proteolytic activity.This conversion is exerted by 2 physiologic plasminogen activators(PAs), tissue-type plasminogen activator (tPA) and urokinase-typeplasminogen activator (uPA). Because unrestrained generation ofthe broad-spectrum protease plasmin may be hazardous to the cells,plasmin activity is tightly controlled at the levels of PAs byplasminogen activator inhibitors (PAI-1 and PAI-2) and at the levelof plasmin bya2-antiplasmin (reviewed by Irigoyen et al1 andCollen and Lijnen2). Plasmin is the major enzyme responsible forthe dissolution of fibrin at both intravascular and extravascularsites. In addition, plasmin has been proposed to lead to matrixdegradation indirectly via activation of matrix metalloproteinases(MMPs). Furthermore, plasmin can activate several latent growthfactors in vitro, including latent transforming growth factor-b andlatent basic fibroblast growth factor, and these activities have beenproposed to be crucial for cell migration and tissue remodeling invivo. Thus, the plasmin system has been shown to play a role in awide variety of physiopathological processes involving extracellu-lar matrix degradation, tissue remodeling, and cell migration,including ovulation, trophoblast invasion, postlactational mam-

mary involution, neurite outgrowth, wound healing, inflammation,angiogenesis, and tumor cell invasion (reviewed by Irigoyen et al1

and Collen and Lijnen2).Extracellular proteolysis takes place during skeletal muscle

formation and pathological muscle regeneration, in which satellitecells play a major role. In response to muscle injury, damagedtissue is infiltrated by fibroblasts, inflammatory cells, and macro-phages.3 Necrotic tissue is removed, revascularization starts, andproliferation of satellite cells is initiated. A number of proteolyticenzymes have been proposed to play a role during muscleregeneration, either in the inflammatory response and/or in themigration of myoblasts across the basal lamina and in their furtherfusion to form the terminal muscle fiber.4,5 Metalloproteinases suchas MMP-2 and MMP-9, meltrin-a, and cathepsin B seem to berequired for myotube formation in vitro.5-8 Moreover, the expres-sion of MMP-2 and MMP-9 has also been reported in thedegeneration-regeneration process of myofibers in vivo.9 Themechanism of MMP activation in most cell types has been found toinvolve a proteolytic activation cascade initiated by uPA. MostMMPs then can be directly activated by cleavage to a lowermolecular weight protein by plasmin.10 Expression of uPA and tPAhas been reported in cell cultures from chicken, mouse, rat, and

From the Centre d’Oncologia Molecular, Institut de Recerca Oncologica,Barcelona, Spain; and the Center for Transgene Technology, FlandersInteruniversity, Leuven, Belgium.

Submitted June 26, 2001; accepted December 11, 2001.

Supported by grants from the European Union 1999/C361/06 (E.S., G.A.,M.S.), DGES PM97-008, Fundacio La Marato-TV3 (F.L., M.P.), and FIS01/1474. M.S. was also supported by the Catalan government (Reincorporaciode Doctors).

Reprints: Pura Munoz-Canoves, Centre d’Oncologia Molecular, Institut deRecerca Oncologica, Aut Castelldefels, km 2.7, 08907 L’Hospitalet de Ll,Barcelona, Spain; e-mail: pmunoz@iro.es.

The publication costs of this article were defrayed in part by page chargepayment. Therefore, and solely to indicate this fact, this article is herebymarked ‘‘advertisement’’ in accordance with 18 U.S.C. section 1734.

© 2002 by The American Society of Hematology

2835BLOOD, 15 APRIL 2002 z VOLUME 99, NUMBER 8

For personal use only. by guest on June 12, 2013. bloodjournal.hematologylibrary.orgFrom

human muscle.11-13 We showed that inhibition of uPA proteolyticactivity with an anti-uPA antibody abrogated migration, fusion, anddifferentiation of murine myoblasts in vitro.14 Moreover, werecently demonstrated that uPA activity is required for efficientskeletal muscle regeneration in vivo.15 However, the role ofplasmin in skeletal myogenesis in vitro and in muscle regenerationin vivo has never been investigated. In this study we show thatinhibition of plasmin activity witha2-antiplasmin abrogates myo-blast fusion and differentiation in vitro. Moreover, using micegenetically deficient in Plg (Plg2/2)16 we show that the regenera-tion capacity of muscle tissue in Plg-deficient mice is severelyimpeded. Because degenerating muscles of Plg-deficient miceshow a dramatic accumulation of fibrin, and because fibrinogendepletion restores the onset of normal muscle regeneration in thesemice, we propose that plasmin-mediated fibrinolysis is required forefficient muscle regeneration in vivo.

Materials and methods

Cell culture

C2C12 murine myoblastic cell line was obtained from the American TypeCulture Collection and maintained in Dulbecco modified Eagle medium(DMEM) containing 10% fetal bovine serum (FBS), designated as growthmedium. Fusion and differentiation were initiated by replacing growthmedium with serum-free DMEM supplemented with 170 nM insulin(Sigma, St Louis, MO), designated as differentiation medium. Whenindicated, DMEM plus insulin was supplemented with 10mM SB203580(Calbiochem, Darmstadt, Germany), 100 nMa2-antiplasmin (kindlyprovided by Dr R. Lijnen, Leuven, Belgium), or 0.2% bovine serumalbumin (BSA) (Sigma). All media were supplemented with 100 U/mLpenicillin and 100mg/mL streptomycin.

Animals

The mice used were 9- to 12-week-old C57Bl16 control strain and thecorresponding Plg-deficient mice.16 For all breading, heterozygous Plg1/2

mice were used. The genotypes of mice were established by polymerasechain reaction (PCR) of tail biopsy DNA. The wild-type (WT) Plg allelewas detected with the primers PLG-1: 59-TCAGCAGGGCAATGTCAC-GG-39and PLG-2: 59-CTCTCTGTCTGCCTTCCATGG-39. These primersgenerate a 450 base pair (bp) product. The targeted Plg allele was detectedwith the primers Neo-1: 59-ATGATTGAACAAGATGGATTGCACG-39and Neo-2: 59-TTCGTCCAGATCATCCTGATCGAC-39. These primersgenerate a 480 bp product. All mice were housed in standard facilities andkept at room temperature with a natural night-day cycle. Prior to manipula-tion, mice were anesthetized by an intraperitoneal injection ofdormicum/droperidol.

Induction of muscle regeneration

Regeneration of skeletal muscle was induced by intramuscular injection of50 mL of 50% glycerol (vol/vol) in the gastrocnemius muscle group of themice, as described by Kawai et al.17 The experiments were performed inright hind-limb muscles, and contralateral intact muscles were used ascontrol. Morphologic and biochemical examinations were performed at 2,4, 7, 10, 30, and 50 days after injury (AI). Four animals were used for eachtime point.

Histologic and immunohistochemical analysis

At selected times, muscles of control and Plg2/2 mice were removed aftercervical dislocation. They were carefully dissected, frozen in isopentane-chilled liquid nitrogen, and stored at280°C prior to sectioning. Transversecryostat sections (10-mm thick) were stained with hematoxylin/eosin (H/E).

Immunohistochemistry was performed using the Vectastain Elite kit (VectorLaboratories, Burlingame, CA) and diaminobenzidine for single-labelingexperiments. The following antibodies were used: mouse monoclonalantibody against myosin developmental-type heavy chain (MHCd) 1:50(Novocastra, Newcastle, United Kingdom) and rabbit antimouse fibrinogen1:1000 (kindly provided by Dr K. Dano, Finsenlab, Denmark). Forimmunofluorescence staining, sections were incubated with rat monoclonalantibodies anti–Mac-1 1:30 (Pharmingen, San Diego, CA), anti–Gr-1 1:30(Pharmingen), and anti-T11 1:50 (Coulter Immunology, Hialeah, FL)conjugated with fluorescein and mounted with Vectashield (Vector Labora-tories). Sections were photographed using an Olympus BX60 microscope(Tokyo, Japan). Control experiments without primary antibody demon-strated that immunofluorescence signals observed were specific (not shown).

RNA analysis

Total RNA was extracted and purified from either C2C12 cells or freshlyisolated muscle tissue according to the procedure of Chomczynski andSacci.18 Five micrograms of total RNA was subjected to Northern analysis,and blots were hybridized with radiolabeled complementary DNA probecorresponding to murine myogenin, as previously described.14,19To normal-ize signal intensity, blots were later rehybridized with a radiolabeled 18Soligonucleotide probe. For reverse transcription (RT)–PCR analysis, 2mgtotal RNA was reverse transcribed using the first-strand complementaryDNA synthesis kit (Pharmacia, Uppsala, Sweden) in a 35-mL reaction.Amplification parameters were denaturation at 95°C for 45 seconds,annealing for 2 minutes at 58°C, and extension at 72°C for 1 minute.Primers for detection of Plg reverse transcriptase product were derived frommurine Plg complementary DNA sequence: forward primer: 59-TCAG-CAGGGAATGTCACGG-39; reverse primer: 59-CTCTCTGTCGCCTTC-CATGG-39. Expected RT-PCR product size was 410 bp.

Preparation of muscle extracts

Gastrocnemius muscles were dissected out and sectioned, weighed, andpounded in an ice-cold Potter tube with 0.1 mM Tris-HCl buffer, pH 7.6,containing 2 mM ethylenediaminetetraacetic acid (EDTA) and 0.4% TritonX-100. The resulting extracts were centrifuged for 20 minutes at 4°C at12 000g,and the supernatants were stored as aliquots at280°C until use.Protein concentration was determined in the supernatants using the Bio Rad(Herts, United Kingdom) protein assay.

Analysis of myogenin protein expression

Tissue extracts from noninjured or injured muscle were prepared asdescribed above, and myogenin expression was analyzed by Westernblotting using a rabbit polyclonal antimyogenin antibody 1:200 (sc-576;Santa Cruz Biotechnology, Santa Cruz, CA). Immunoblots were performedand developed with the ECL detection system (Amersham, Buckingham-shire, United Kingdom).

Plasmin kinetics

Plasmin content in 20mL conditioned media (5-fold concentrated, usingCentricon, Pall Filtron, Ann Arbor, MI) from C2C12 cells in DMEM orDMEM supplemented with insulin was measured using the S-2251–basedassay, which was done in triplicate in microtitration plates at 37°C andcalibrated with purified plasmin (Chromogenix, Milan, Italy). The chromo-genic substrate selective for plasmin, S-2251 (Chromogenix), was used tofollow the initial rate of Plg activation by measuring p-nitroanilinegeneration. Conditioned media were mixed with a buffer containing 0.1 MTris-HCl and 2 mM EDTA, pH7.6, and 1.6 mM S-2251 as substrate. Thegeneration of plasmin was detected by measuring the p-nitroaniline releasefrom the substrate as indicated above.

Zymography

Sixty micrograms of muscle extracts or 40mL C2C12 conditioned mediawere size-fractionated on a 10% nonreducing sodium dodecyl sulfate (SDS)

2836 SUELVES et al BLOOD, 15 APRIL 2002 z VOLUME 99, NUMBER 8

For personal use only. by guest on June 12, 2013. bloodjournal.hematologylibrary.orgFrom

acrylamide gel, which was washed for 30 minutes in 2.5% TritonX-100/phosphate-buffered saline and for 30 minutes in distilled water. Thegel was subsequently placed in contact with a casein gel containing 2%(wt/vol) nonfat dry milk, 0.25 mM Tris-HCl (pH 7.6), 1% (wt/vol) agarose,0.253 phosphate-buffered saline, and 15mg/mL Plg (Chromogenix) andincubated in a humid chamber at 37°C until caseinolytic bands werevisualized and photographed.

Systemic defibrinogenation

Plg-deficient mice were anesthetized, and 14-day mini-osmotic pumps(model 1002, Alza, Palo Alto, CA), filled with a buffered solution of 500U/mL ancrod (Sigma Chemical), were implanted subcutaneously into theirbacks (one minipump per animal). The insertion sites were then closed bysutures. The pumps deliver 0.25mL/h, so the mice received 3 U ancrod/d. Incontrol animals, saline-filled minipumps were implanted. On day 3 ofancrod or saline infusion, injury was induced by intramuscular injection of50% glycerol. Ten days after injury, mice were killed and gastrocnemiusmuscles were dissected, frozen, and analyzed by H/E staining. Fibrinogenlevels in citrated blood from ancrod- or saline-treated mice were analyzedby SDS–polyacrylamide gel electrophoresis (PAGE) (6% polyacrylamidegel) followed by Western blotting using an antimouse fibrinogenantibody 1:1000.

Statistical analysis

The Studentt test was used to determine whether there were significant(P , .05) differences between WT and Plg-deficient mice.

Results

C2C12 muscle cells produce uPA-dependent plasmin activity

Previous studies have reported that myogenic cells express Plgactivators in vitro.11,12,14However, the presence of the downstreamprotease plasmin in muscle cells has never been documented. Thus,we investigated whether plasmin activity was generated by C2C12cells (a myoblast cell line derived from murine satellite cells),which have been extensively used as a model for myogenesis invitro.20 First, we demonstrated by RT-PCR that Plg messengerRNA was synthesized in C2C12 cells (Figure 1A). Using a specificchromogenic substrate for the activity of plasmin, we observed thatplasmin activity was detected both in conditioned medium ofproliferating preconfluent C2C12 myoblasts (lane 1) (Figure 1B)and in conditioned medium of fusing myotubes (lane 2) (Figure1B). To analyze whether the generation of plasmin by C2C12 cellswas uPA- or tPA-dependent, we analyzed the activity of both PAs inthese cells by zymography. This assay allowed clear distinctionbetween uPA and tPA, which migrated at 45 and 72 kd, respec-tively. As shown in Figure 1C, uPA proteolytic activity was clearlyobserved in C2C12-conditioned medium of both myoblasts (lane 1)and myotubes (lane 2), whereas tPA activity was practicallyundetectable. These results suggested that the generation of plas-min in C2C12 muscle cells was mediated mainly by uPA.

Inhibition of plasmin activity reduces myotube formation andmuscle differentiation in vitro

To investigate the role of plasmin in C2C12 myogenesis, weanalyzed whether inhibition of plasmin activity affected myoblastfusion and differentiation in vitro. C2C12 cells cultured in DMEMplus 10% FBS were transferred to DMEM containing insulin for 3days, to induce myoblast fusion and differentiation, in the absence

or presence of the physiologic plasmin inhibitor,a2-antiplasmin,and myotube formation and differentiation were analyzed conse-quently. As shown in Figure 2A,B, in the absence ofa2-antiplasmin, C2C12 cells fused extensively; in contrast, fewermyotubes formed in cultures that had been treated witha2-antiplasmin; as an experimental control, we observed that SB203580(a specific inhibitor of p38 mitogen–activated protein kinase)significantly reduced myoblast fusion, as previously reported,21,22

while no decrease in myotube formation was observed in thepresence of BSA. The effect ofa2-antiplasmin on myotubeformation was determined microscopically by counting the fractionof nuclei present in myotubes. The results represented in the graph(Figure 2A) show thata2-antiplasmin reduced myotube formationby 45%. SB203580 reduced myotube formation by 80%, whiletreatment of cells with BSA had practically no effect. To establishwhether the inhibition of plasmin activity also affected myogenicdifferentiation, RNA was obtained from cells treated or not witha2-antiplasmin, SB203580, or BSA and analyzed for the expres-sion of the myogenic differentiation marker myogenin. As shownin Figure 2C, myogenin transcripts were absent in preconfluentmyoblasts grown in DMEM plus 10% FBS (lane 1), while theiraccumulation was induced when C2C12 cells were switched toDMEM plus insulin for 3 days, coinciding with extensive myotubeformation (lane 2). When differentiating C2C12 cells were incu-bated witha2-antiplasmin (lane 3) or SB203580 (lane 4) for 3

Figure 1. C2C12 myoblasts produce uPA-dependent plasmin activity. (A) C2C12muscle cells express Plg. RT-PCR was performed to detect the expression of Plg inmurine liver (positive control, lane 2) and differentiating C2C12 cells (lane 3). Lane 1corresponds to PCR-negative control. Expected RT-PCR product size was 410 bp.(B) Plasmin activity is produced by C2C12 myoblasts. Histograms show plasminspecific activity (SA) (expressed as mOD/min/mg protein) in conditioned medium ofC2C12 cells: preconfluent myoblasts (lane 1) and fusing myotubes (lane 2). Noplasmin activity is detected in DMEM medium (lane 3), used as a negative control.The assays were performed in 0.1 M Tris-HCl and 2 mM EDTA, pH 7.6, containing thechromogenic substrate S-2251. (C) C2C12 cells produce uPA proteolytic activity.Conditioned media from preconfluent myoblasts (lane 1) and fusing myotubes (lane2) cells were analyzed, after SDS-PAGE, for uPA and tPA activity, by PA zymography.The 45-kd approximate molecular mass species indicated by an arrow correspondsto murine uPA and was calculated according to standard molecular mass markerselectrophoresed in an adjacent lane and stained with Coomassie blue. The photo-graphs were taken after overnight incubation at 4°C and 4 hours at 37°C.

PLASMIN IS REQUIRED FOR MYOGENESIS 2837BLOOD, 15 APRIL 2002 z VOLUME 99, NUMBER 8

For personal use only. by guest on June 12, 2013. bloodjournal.hematologylibrary.orgFrom

days, the levels of myogenin messenger RNA were significantlydecreased; in contrast, no alteration in the levels of these transcriptswas observed when BSA was added to the cells (lane 5).Furthermore, expression of myogenin was already detected 18hours following incubation of C2C12 cells in differentiationmedium, at a stage when no myotubes had been formed (Figure 2D,lane 1), while in the presence ofa2-antiplasmin or SB203580 thelevels of myogenin were reduced (lanes 2 and 3, respectively).Altogether, these results indicate that plasmin activity is requiredfor complete myoblast fusion and differentiation in vitro.

Plasmin production in regenerating muscle is dependenton uPA activity

We next investigated the functional relevance of plasmin duringskeletal muscle regeneration in vivo. Muscle regeneration wasinduced in mice by intramuscular injection of 50% glycerol.17

Using the chromogenic substrate S-2251 for the activity ofplasmin, we had shown previously that plasmin generation wasincreased during skeletal muscle regeneration in WT mice, witha peak of activity occurring 2 to 4 days AI and decreasing byday 10.15

To investigate whether the generation of plasmin from Plg isuPA- or tPA-dependent, we analyzed uPA and tPA activities inregenerating muscle tissues at different times AI. Tissue extractswere prepared from noninjured muscle (0 days AI); from 2-, 4-, and7-day–injured muscle (2, 4, and 7 days AI); and analyzed by

zymography (Figure 3). As mentioned above, this assay allowsclear distinction between uPA and tPA, which migrate at 45 and 72kd, respectively. No uPA activity was detectable in tissue extractsfrom noninjured muscle. However, it was induced in the regenerat-ing muscle samples, reaching a peak at 4 days AI, as demonstratedby the appearance of a casein degradation band of 45 kd,corresponding to murine uPA active enzyme. Under the sameconditions, tPA activity was undetectable in all muscle extractsanalyzed, while purified human uPA (55 kd), utilized as a controlfor activity (C), was detected in an adjacent lane. These resultssuggested that uPA, rather than tPA, might be involved in Plgactivation during skeletal muscle regeneration in vivo.

Expression of myogenin is reduced in regenerating muscletissue of Plg-deficient mice

To evaluate the consequences of loss of plasmin in the skeletalmuscle regeneration process, we induced muscle injury in WT andPlg2/2 mice by glycerol injection, and the expression myogeninwas analyzed because the expression of this transcription factor isknown to be induced in satellite cells during muscle regeneration invivo. RNA and protein lysates were prepared from muscles on thefourth day after injury (W, wound) as well as from noninjuredmuscles (C, control). As shown by Northern blotting, myogeninRNA levels were very low in quiescent muscle from WT mice andwere dramatically induced in injured muscle (Figure 4A, compareC and W in WT), indicating that the regeneration process wasproceeding normally. However, myogenin transcript levels wereinduced to a lesser extent in Plg-deficient mice than in WT miceafter injury (Figure 4A, compare W in WT and Plg2/2), while the18S RNA level was comparable in all lanes. Similarly, followingmuscle injury, myogenin protein levels were induced to a greaterextent in WT mice than in Plg-deficient mice (Figure 4B, compareW in WT and Plg2/2). These results indicated that Plg is requiredfor efficient expression of muscle regeneration specific geneproducts. As a control for the the Plg2/2 genotype, PCR genotypingof WT, Plg2/2, and Plg1/2 mice was performed using specificprimers. As shown in Figure 4C, PCR products of 450 and 480 bp,corresponding to WT Plg allele and targeted allele, respectively,were obtained.

Figure 3. The proteolytic activity of uPA is increased in regenerating muscleafter injury. Forty micrograms of proteins from muscle extracts of noninjured (0 daysAI) and injured skeletal muscle (2, 4, and 7 days AI). were analyzed after SDS-PAGEby PA zymography. Two nanograms of purified human uPA (55 kd) were utilized as acontrol for activity (C). The 45-kd approximate molecular mass species correspondsto murine uPA and was calculated according to standard molecular mass markerselectrophoresed in an adjacent lane and stained with Coomassie blue. The photo-graphs were taken after overnight incubation at 4°C and 4 hours at 37°C.

Figure 2. Inhibition of plasmin activity with a2-antiplasmin decreases myotubeformation and differentiation. C2C12 myoblasts were grown in DMEM plus 10%FBS and then switched to DMEM supplemented with insulin for 3 days (A, B, C), toinduce myotube formation and differentiation, or 18 hours (D), to induce onlymyogenic differentiation, in the presence or absence of a2-antiplasmin, SB203580,or BSA. (A) The a2-antiplasmin decreases myotube formation. Quantitative effects ofplasmin inhibition on myoblast fusion. Near-confluent C2C12 cells were switched toDMEM plus insulin to promote fusion in the absence (E) or presence of a2-antiplasmin (‚), SB203580 (f), or BSA (M). Fusion is represented as percentage ofnuclei in myotubes. Values are mean of 3 independent experiments. (B) Lower-powerviews of representative fields. Preconfluent cells in DMEM/10% FBS (i), confluentcells in DMEM/insulin (ii), confluent cells in DMEM/insulin plus a2-antiplasmin (iii).Magnifications are 3 100. (C) The a2-antiplasmin decreases myogenic differentia-tion. Northern blot analysis of 5 mg total RNA from C2C12 cells grown as myoblasts inDMEM plus 10% FBS (lane 1) and then switched to DMEM plus insulin for 3 days toinduce myoblast differentiation and fusion, in the absence (lane 2) or presence ofa2-antiplasmin (lane 3), SB203580 (lane 4), or BSA (lane 5). The probe for eachhybridization is indicated. (D) Inhibition of plasmin activity with a2-antiplasmindecreases myogenin expression prior the onset of myotube formation. Northern blotanalysis of 10 mg total RNA from C2C12 cells grown as myoblasts in DMEM plus 10%FBS and the switched for 18 hours to DMEM plus 170 nM insulin to induce myogenicdifferentiation, in the absence (lane 1) or presence of 100 nM a2-antiplasmin (lane 2)or 10 mM SB203580 (lane 3). The probe for each hybridization is indicated.

2838 SUELVES et al BLOOD, 15 APRIL 2002 z VOLUME 99, NUMBER 8

For personal use only. by guest on June 12, 2013. bloodjournal.hematologylibrary.orgFrom

Plg-deficient mice show a severe regeneration defect withenhanced fibrosis and myotube degeneration

To analyze the functional significance of the increased plasmin activityin skeletal muscle after damage, we performed glycerol-induced muscleinjury in WT mice and in Plg2/2 mice, and we analyzed comparativelythe histopathological changes induced by the lesion in both types ofmice. Skeletal muscle of Plg-deficient mice showed a dramatic regenera-tion defect after glycerol-induced injury, whereas regeneration in WTmice proceeded normally (Figure 5A). This regeneration defect inPlg2/2 mice was apparent by 4 days AI but was most striking 10 to 30days after injury. Analysis of H/E-stained cross-sections of WT andPlg-deficient mice 2 days after injury showed that muscles of all micewere edematous, presenting fibrotic infiltrates within the enlargedintercellular space separating the necrotic myofibers. Cross-sections ofmuscles 4 days after injury already showed features of ongoingregeneration in WT mice, with many new myofibers characterized bysmall size and single nuclei, and with a reduction in fibrotic infiltrates(Figure 5A). In contrast, in Plg-deficient mice, at 4 days AI the muscleappeared edematous and no new uninucleated, small myofibers weredetected yet. In WT mice, at 7 days AI most injured fibers regeneratedinto groups of centrally nucleated myotubes, an indication of advancedregeneration, and a very small number of necrotic fibers could beobserved; in addition, the size of newly formed myofibers had increased

in the WT animals. In contrast, in Plg-deficient mice, the muscleappeared necrotic, showing still extensive fibrosis 7 days after damage.Ten days after injury, the muscle of WT mice was completelyregenerated; no sign of previous damage was detectable in WT mice,except for the presence of centrally located nuclei inside the regeneratedfibers (Figure 5A). In Plg-deficient mice, however, a high number ofdegenerated myotubes were still visible. Thirty days after injury, thelesion was no longer noticeable in WT mice, except for the centralmyonuclei. In Plg2/2 mice, however, the muscle presented extensivefibrosis with high numbers of degenerated myotubes. Notably, extensivemuscle degeneration was still observed in Plg-deficient mice 50 daysafter injury, suggesting that regeneration is not only retarded, butdramatically impaired, in mice deficient in Plg (see Figure 5A).

To confirm further the muscular regeneration defect in Plg2/2

mice, we performed an immunohistochemical analysis of musclesections of WT and Plg-deficient mice at 4 and 10 days after injuryusing a monoclonal antibody against MHCd. The expression of thisMHC isoform occurs exclusively in embryonic and early neonatallife and upon muscle regeneration in adult life. As shown in Figure5B, the regenerating myofibers of both WT and Plg-deficient miceexpressed MHCd 4 days AI; in contrast, at 10 days AI MHCdexpression was detected in Plg2/2 mice but not in WT mice. Theseresults clearly indicated that the muscles of Plg-deficient mice

Figure 4. Distinct up-regulation of myogenin in skeletal muscle ofWT and Plg-deficient mice after freeze-crush injury. RNA andtissue extracts were obtained from the gastrocnemious muscle of WTand Plg2/2 mice at 0 days AI (C, control) or 4 days AI (W, wound).(A) Analysis of myogenin messenger RNA in regenerating muscle ofWT and Plg2/2 mice. Northern analysis was performed to detectexpression of myogenin and 18S. Myogenin transcript levels wereinduced in regenerating muscle of WT mice. Expression of myogeninwas lower in regenerating muscle of Plg-deficient mice compared withregenerating muscle of WT mice, whereas levels of 18S were similarin all lanes. (B) Analysis of myogenin protein in regenerating muscle ofWT and Plg2/2 mice. Western analysis was performed to detectexpression of myogenin using an antimyogenin antibody. PonceauRed staining was performed to check for equal protein loading.Myogenin was induced in both WT and Plg-deficient mice, althoughinduction was greater in the former mice. (C) PCR genotyping ofPlg1/1, Plg2/2, and Plg1/2 mice. PCR products of 450 and 480 bpcorrespond to WT Plg allele and targeted allele, respectively.

PLASMIN IS REQUIRED FOR MYOGENESIS 2839BLOOD, 15 APRIL 2002 z VOLUME 99, NUMBER 8

For personal use only. by guest on June 12, 2013. bloodjournal.hematologylibrary.orgFrom

present an ongoing regeneration/degeneration process at stageswhen muscle regeneration is almost complete in WT mice.

Plg deficiency alters the inflammatory response aftermuscle injury

Circulating inflammatory cells are recruited after injury to skeletalmuscle during the inflammatory phase.3 While neutrophils have beendetected in the early hours following muscle injury, maximal levels ofmacrophages and T lymphocytes appear later during the muscle injuryresponse. It is widely held that Plg system plays a role in inflammationthrough plasmin-mediated directional cell migration. To directly charac-terize plasmin’s involvement in the inflammatory response followingmuscle injury, the presence of neutrophils, macrophages, and T lympho-cytes in normal and in regenerating muscle of WT and Plg-deficientmice was analyzed by immunohistochemistry using antibodies againstGr-1, Mac-1, and T11, well-characterized neutrophil, macrophage, andT-lymphocyte markers, respectively. No Gr-11 cells were detected in

noninjured muscle sections of WT or Plg-deficient mice (Figure 6A, 0days AI). Twelve hours after injury, there was a similar increase inGr-1–expressing cells at the sites of injury in both types of animals(Figure 6A, 0.5 days AI,P, .05). Only a few resident Mac-11 cellswere detected in the intact muscle sections of WT and Plg2/2 mice(Figure 6B, 0 days AI). Two days AI, the number of Mac-1–expressingcells augmented at the injury site of both types of animals; however, thenumber of Mac-11 cells in Plg-deficient mice was reduced to almost50% with respect to WT mice (Figure 6B, 2 days AI,P, .05).Similarly, 2-days AI the number of T11-expressing cells increased withrespect to resting muscle in both WT and Plg-deficient mice; however,the number of T11-expressing cells at the injury site in muscles ofPlg-deficient mice was 35% lower than in WT mice (Figure 6C, 2 daysAI, P , .05). Quantification of these data are represented in Figure6D. These results suggested that the recruitment of both macro-phages and T lymphocytes to the injured muscle was reduced in theabsence of plasmin.

Figure 5. Skeletal muscle regeneration is defective in Plg-deficient mice. (A) Glycerol-induced injuries result in impaired skeletal muscle regeneration of Plg-deficientmice. Frozen sections of muscles from WT and Plg-deficient mice were stained with H/E after 2, 4, 7, 10, 30, and 50 days AI, as indicated. Contralateral control muscles werealso stained with H/E (0 days AI). In WT mice, advanced regeneration is visible after 4 days, and regeneration is complete after 10 days. In Plg-deficient mice, a regenerationdefect is already visible after 4 days AI but is most striking after 10 days. In WT mice, virtually no sign of the previous injury was detectable after 30 and 50 days, except for thecentrally located myonuclei. Original magnifications 3 400. (B) Persistence of MHCd-positive myofibers 10 days after glycerol injury in Plg-deficient mice. Cryostat frozensections from WT and Plg-deficient mice were reacted with a monoclonal antibody against MHCd at 4 and 10 days AI, as indicated. Large numbers of MHCd-expressingmyofibers are visible in regenerating muscle of Plg-deficient mice 4 and 10 days after intramuscular glycerol injection. In WT mice, MHCd-positive fibers are observed 4 daysAI, whereas no MHCd-expression is detected 10 days AI. Original magnification 3 400.

2840 SUELVES et al BLOOD, 15 APRIL 2002 z VOLUME 99, NUMBER 8

For personal use only. by guest on June 12, 2013. bloodjournal.hematologylibrary.orgFrom

Systemic fibrinogen depletion restores muscle regenerationin Plg-deficient mice

The classical role of plasmin is the degradation of fibrin depositsor fibrinolysis (reviewed by Lijnen and Collen23). In a previousstudy, we demonstrated an accumulation of fibrin followingmuscle injury in Plg-deficient mice, but not in WT mice,correlating with the persistence of muscular degeneration fea-tures in Plg2/2 mice after muscular injury.15 Depletion ofcirculating fibrinogen can be achieved by administration of theviper venom ancrod.15,24 Ancrod- or saline-delivering osmoticpumps (3 U/d) were implanted in Plg2/2 mice during 3 daysbefore muscle injury and then throughout the experimentalperiod (up to 10 days AI). Western blot analysis using anantifibrinogen antibody demonstrated the reduction of plasmafibrinogen upon ancrod administration, without any effect onsurvival (Figure 7A, compare ancrod and saline). Furthermore,signs of improved muscular regeneration were detectable inancrod-treated Plg-deficient mice 10 days after injury, withcentrally located nuclei inside the regenerated fibers. In contrast,histologic features of muscle degeneration persisted in saline-treated Plg-deficient mice at the same stage AI, including highnumbers of degenerated myotubes and fibrosis in the intercellu-lar space throughout the damaged muscle (Figure 7B). Theextent of muscle regeneration of ancrod-treated Plg-deficientmice 10 days AI resembled the regeneration status of a WT

mouse at the same time AI. This observation supports the idea ofa potential pathogenic role of fibrin(ogen) accumulation duringmuscle regeneration in Plg-deficient mice.

DiscussionThe aim of our work was to investigate the involvement of thebroad-spectrum serine protease plasmin in cellular and tissularremodeling occurring during skeletal myogenesis in vitro andskeletal muscle regeneration in vivo. Our results clearly show thatinhibition of plasmin activity bya2-antiplasmin reduced the extentof myoblast fusion and differentiation in vitro. Moreover, micelacking Plg showed a pronounced regeneration defect after experi-mentally induced injury of muscle, suggesting a protective role forPlg in the muscular regeneration process. We recently demon-strated persistent muscle degeneration in uPA-deficient but not intPA-deficient mice.15 Altogether, this indicates that between the 2pathways of Plg activation (uPA- or tPA-mediated), uPA-mediatedPlg activation is the major one in the muscle regeneration process.This conclusion is supported by zymographic analysis of regenerat-ing muscles of WT mice, which only showed uPA activity.Interestingly, in C2C12 murine myoblasts, uPA activity was alsopredominant. Finally, the similar phenotypes of uPA- and Plg-deficient mice indicates, in this experimental muscle regenerationmodel, the predominance of uPA-dependent plasmin effects.

Figure 6. Immunohistochemical demonstration ofthe presence of inflammatory cells in muscles of WTand Plg-deficient mice after glycerol-induced injury.Detection of neutrophils (Gr-11) (A), macrophages (Mac-11) (B), and T cells (T111) (C) in transverse sections ofcontrol and injured skeletal muscles of WT and Plg-deficient mice at the indicated times AI. (D) Quantitativeanalysis of Gr-11, Mac-11, and T111 cells in regeneratingmuscles of WT and Plg-deficient mice. Gr-11, Mac-11,and T111 cells were counted in 10 random fields persection, and data from 12 sections per muscle werepooled. The mean values for both limbs were to provide amean value for each animal (3 animals were studied ateach time point). Histograms show the percentage ofreduction of Mac-11 and T111 cells (50% and 30%,respectively) in Plg2/2 mice (gray bars) with respect toWT mice (black bars); statistical significant (P , .05) isdenoted. Gr-1–expressing cells were maximal 12 hoursAI in both WT and Plg-deficient mice. Mac-1– andT11-expressing cells increased 2 days AI in muscle ofWT and Plg-deficient mice; however, the number ofMac-11 and T111 cells was lower in Plg2/2 mousemuscle. Original magnifications 3 400.

PLASMIN IS REQUIRED FOR MYOGENESIS 2841BLOOD, 15 APRIL 2002 z VOLUME 99, NUMBER 8

For personal use only. by guest on June 12, 2013. bloodjournal.hematologylibrary.orgFrom

Extravascular fibrin deposition is a key feature in pathologiescharacterized by inflammation and tissue repair, including impairedskin wound healing25 and glomerulonephritis.26 In these 2 situa-tions, failure to remove plasmin was attributable to reduced uPA- ortPA-mediated fibrinolysis, respectively. In experimentally inducedmuscle degeneration, we also found that uPA deficiency led toincreased levels of muscular fibrin, indicating the importance ofuPA in extravascular fibrin clearance.15 In the present study, usingPlg-deficient mice we have demonstrated the importance ofplasmin for fibrin clearance following muscle injury. Fibrin accu-mulation in the extracellular basal membrane may have deleteriouseffects, such as the impediment of normal nutrition to the muscletissue. The hypothesis that increased fibrin levels may contribute toimpaired muscular regeneration was further explored by systemicfibrinogen depletion of Plg-deficient mice. Administration ofancrod reduced plasma fibrinogen levels in Plg-deficient mice,resulting in an almost complete restoration of the normal muscleregeneration process. These results clearly show that fibrin(o-gen) accumulation has a pathogenic role in sustaining muscledegeneration.

Our finding that loss of plasmin activity impedes muscleregeneration raises the question of how this proteolytic activity isinvolved in tissue repair. Inflammation is a process frequentlyassociated with tissue repair, because degenerating tissues areinvaded by inflammatory cells. Our study showed that in responseto glycerol-induced muscle injury, neutrophils, macrophages, and Tlymphocytes accumulated near the injury site in the first days AI.Similar results in macrophage recruitment were reported in a studyof regeneration following bupivacaine-induced muscle injury in therat27 and following crush injury in mouse.3 We have observed thatmice with a specific deficit in Plg show a reduced staining forMac-11 and T111 cells 48 hours AI, indicating that the number ofmacrophages and T lymphocytes reaching the injury site is reducedin the absence of Plg. Recently, deficient macrophage recruitmentwas also reported by our group in regenerating muscle of uPA-deficient mice.15 This suggests that uPA/plasmin activity may havea profound effect on inflammation and inflammation-related muscu-lar disease. Previous studies with uPA-deficient mice demonstratedthat uPA is required for the pulmonary inflammatory response toCryptococcus neoformans,because a lack of uPA resulted ininadequate macrophage recruitment, uncontrolled infection, anddeath.28 Similarly, monocyte and lymphocyte recruitment wassignificantly diminished in Plg-deficient mice after thioglycollate-

induced acute peritoneal inflammation.29Additionally, componentsof the Plg system may regulate the expression or/and activity ofcytokines involved in inflammatory processes. For example, endo-genously produced uPA can amplify tumor necrosis factor-aneosynthesis by mononuclear phagocytes, representing a novelmechanism by which a phagocyte-derived protease contributes togenerating proinflammatory signals.30 In addition, plasmin hasbeen shown to release macrophage-derived interleukin-1 and toactivate transforming growth factor-b.31,32 It is therefore temptingto speculate that uPA/plasmin might be playing a similar role in theinflammation caused by muscle injury. Thus, the reduced presenceof macrophages and T lymphocytes in the injured muscles ofPlg-deficient mice might be due either to a reduction in themigration capacity of inflammatory cells devoid of plasmin activityor to a reduced potential of these cells to traverse fibrin-richmatrices. Furthermore, it has been shown that activated macro-phages (equivalent to those that accumulate at the site of muscledamage) produce soluble factors (fibroblast growth factor andplatelet-derived growth factor), which are highly chemoattractantand also mitogenic for muscle precursor cells.33 Thus, the activatedmacrophages that accumulate in response to muscle damage willnot only phagocytose necrotic tissue but also facilitate the repair ofdamaged myofibers. These processes will be altered if macro-phages do not accumulate at the injury site, providing a potentialexplanation for the persistent muscle degeneration observed inPlg-deficient mice.

The hypothesis that, in general, Plg activation facilitates cellularpenetration of fibrin-containing matrices fits with a putativescenario occurring in normal skeletal muscle regeneration, withlocal conversion of Plg to plasmin and subsequent fibrin degrada-tion. Plasmin is probably one member of a team of carefullyregulated and specialized matrix degrading enzymes, includingserineproteases, metalloproteases, and other classes of proteases,which together serve in matrix remodeling and cellular reorganiza-tion of wound fields. Our data suggest that plasmin is particularlyuseful in fibrin solubilization and, without it, cellular reorganiza-tion of fibrin reach matrices is severely impeded. We conclude thatthe uPA/Plg system is required for normal regeneration andresolution of muscle damage in mice. The reduced presence ofmacrophages at the injury site suggests that a key factor slowingmuscle repair in Plg-deficient animals may be the diminishedability of responding macrophages to proteolytically dissect theirway through the extracellular matrix to the injury site. In this

Figure 7. Effect of systemic fibrinogen depletion onmuscle regeneration of Plg-deficient mice. Plg-deficient mice were implanted 3 days before glycerol-induced muscle degeneration with ancrod-filled mini-osmotic pumps (3U/d over 13 days) (ancrod) or withbuffered-filled pumps (saline). (A) Ancrod treatment re-duces fibrinogen in blood of Plg-deficient mice. Fortymicrograms of sodium citrate–treated blood obtainedfrom saline- or ancrod-infused mice was analyzed in a 6%polyacrylamide gel by Western blotting using a polyclonalantimouse fibrinogen antibody (1:1000). The 340-kd ap-proximate molecular mass species indicated by an arrowcorresponds to murine fibrinogen and was calculatedaccording to standard molecular mass markers electro-phoresed in an adjacent lane. (B) Fibrinogen depletionrestores muscle regeneration in Plg-deficient mice. Cryo-sections from saline- and ancrod-treated Plg-deficientmice were stained with H/E 10 days AI. While well-advanced regeneration is visible in ancrod-treated Plg2/2

mice, a regeneration defect is still observed in saline-treated mice 10 days after injury. Original magnification 3400.

2842 SUELVES et al BLOOD, 15 APRIL 2002 z VOLUME 99, NUMBER 8

For personal use only. by guest on June 12, 2013. bloodjournal.hematologylibrary.orgFrom

model, one major matrix component within damaged areas thatmay represent a particular impediment to inflammatory cellmigration in the absence of Plg is fibrin. However, based on theexisting data, we cannot exclude that the plasmin deficiency mayimpede cell migration because of the lost contribution of plasmin todegrade other matrix components or the lack of other key matrixproteinases or of growth factors. Our results have shown thatplasmin is required for myogenesis in vitro, because inhibition ofplasmin activity witha2-antiplasmin abrogated fusion and differen-tiation of C2C12 cells (Figures 1 and 2). Although the exactmechanism of this effect needs further examination, it seems likelythat plasmin may be required for myoblast migration and alignmentprior to myoblast fusion. Metalloproteinases such as MMP-2 andMMP-9, meltrin-a, and cathepsin B are also required for myotubeformation in vitro,5-8 and MMP-2 and MMP-9 expression has beenreported in a cardiotoxin-induced degeneration-regeneration pro-cess of myofibers in vivo.9 Moreover, we have also found inductionof both MMP-2 and MMP-9 following glycerol-induced skeletalmuscle regeneration in WT mice (R.L.-A. and P.M.-C, unpublishedresults, 2001). Because most MMPs then can be directly activatedby cleavage to a lower molecular weight protein by plasmin,10 it istempting to speculate that activation of MMP-2 and MMP-9 inregenerating skeletal muscle might also be mediated by plasminproteolysis. Apart from its effects on extracellular matrix metallo-proteases, uPA/plasmin can cleave and activate latent forms ofgrowth/angiogenic factors such as transforming growth factor-band hepatocyte growth factor/scatter factor.34,35 Both factors areexpressed within injured muscle and are believed to have importantmuscle regeneration effects by promoting the activation of quies-cent satellite cells after injury in vivo.36,37Therefore, modulation of

plasmin proteolytic activity may indirectly influence cell recruit-ment and the growth and differentiation of cellular constituents inregenerating muscle, although this role remains to be demonstratedboth in vitro and in vivo. In line with this, a defect in thedifferentiation of satellite cells of MyoD-deficient mice has beendescribed38 and, as a consequence, the skeletal muscle regenerationcapacity of these mice is compromised. Interestingly, the pheno-types of Plg-deficient mice (this study) and MyoD-deficient miceappear to overlap partially.

In conclusion, our results demonstrate that plasmin fulfills abeneficial role in skeletal muscle regeneration, mainly throughuPA-mediated fibrinolytic activity. Compounds aimed at decreas-ing fibrin levels in muscle may be clinically useful in dystrophicmuscular therapies. Future experiments will be directed towardfurther definition of the benefit of defibrinogenating, anticoagulant,or fibrinolytic agents in dystrophinopathies. This report constitutesthe first demonstration for a role of plasmin in skeletal myogenesisin vitro and muscle regeneration in vivo. Rescue experiments willshow whether exogenously supplied plasmin can compensate forthe mutation. If this is the case, PAIs might prove to be useful forthe treatment of muscle injuries or muscle dystrophies, particularlyif uPA/plasmin evokes the response preferentially in muscle.

Acknowledgments

We are very grateful to Dr Keld Dano for the antifibrinogenantibody and Dr R. Lijnen fora2-antiplasmin. The helpfulassistance of E. Ramırez is acknowledged.

References

1. Irigoyen JP, Munoz-Canoves P, Montero L, Kozic-zak M, Nagamine Y. The plasminogen activatorsystem: biology and regulation. Cell Mol Life Sci.1999;56:104-132.

2. Collen D, Lijnen HR. Basic and clinical aspects offibrinolysis and thrombolysis. Blood. 1991;78:3114-3124.

3. Pimorady-Esfahani A, Grounds M, McMenaminPG. Macrophages and dendritic cells in normaland regenerating murine skeletal muscle. MuscleNerve. 1997;20:158-166.

4. Hughes SM, Blau HM. Migration of myoblastsacross basal lamina during skeletal muscle devel-opment. Nature. 1990;345:350-353.

5. Couch CB, Strittmatter WJ. Rat myoblast fusionrequires metalloendoprotease activity. Cell. 1983;32:257-265.

6. Guerin CW, Holland PC. Synthesis and secretionof matrix-degrading metalloproteases by humanskeletal muscle satellite cells. Dev Dyn. 1995;202:91-99.

7. Yagami-Hiromasa T, Sato T, Kurisaki T, Kamijo K,Nabeshima Y-I, Fujisawa-Sehara A. A metallopro-tease-disintegrin participating in myoblast fusion.Nature. 1995;377:652-656.

8. Gogos JA, Thompson R, Lowry W, Sloane BF,Weintraub H, Horwitz M. Gene trapping in differ-entiating cell lines: regulation of the lysosomalprotease cathepsin B in skeletal myoblast growthand fusion. J Cell Biol. 1996;134:837-847.

9. Kherif S, Lafuma C, Dehaupas M, et al. Expres-sion of matrix metalloproteinases 2 and 9 in re-generating skeletal muscle: a study in experimen-tally injured and mdx muscles. Dev Biol. 1999;205:158-170.

10. Lijnen HR, Van Hoef B, Lupu F, Moons L, Carme-liet P, Collen D. Function of the plasminogen/plas-

min and matrix metalloproteinase systems aftervascular injury in mice with targeted inactivationof fibrinolytic system genes. Arterioscler ThrombVasc Biol. 1998;18:135-145.

11. Miskin R, Easton TG, Reich E. Plasminogen acti-vator in chick embryo muscle cells: induction ofenzyme by RSV, PMA and retinoic acid. Cell.1978;15:1301-1312.

12. Quax PHA, Frisdal E, Pedersen N, et al. Modula-tion of activities and RNA level of the componentsof the plasminogen activation system during fu-sion of human myogenic satellite cells in vitro.Dev Biol. 1992;151:166-175.

13. Barlovatz-Meimon G, Frisdal E, Hantai D, Angles-Cano E, Gautron J. Slow and fast rat skeletalmuscles differ in their plasminogen activator ac-tivities. Eur J Cell Biol. 1990;52:157-162.

14. Munoz-Canoves P, Miralles F, Baiget M, FelezJ. Inhibition of urokinase-type plasminogen acti-vator (uPA) abrogates myogenesis in vitro.Thromb Haemost. 1997;77:526-534.

15. Lluis F, Roma J, Suelves M, et al. Urokinase-dependent plasminogen activation is required forefficient skeletal muscle regeneration in vivo.Blood. 2001;97:1703-1711.

16. Ploplis V, Carmeliet P, Vazirzadeh S, et al. Effectsof disruption of the plasminogen gene on throm-bosis, growth and health in mice. Circulation.1995;92:2585-2593.

17. Kawai H, Nishino H, Kusaka K, Naruo T, TamakiY, Iwasa M. Experimental glycerol myopathy: ahistological study. Acta Neuropathol. 1990;80:192-197.

18. Chomczynski P, Sacchi N. Single-step method ofRNA isolation by acid guanidinium thiocyanate-phenol-chlorophorm extraction. Anal Biochem.1987;162:156-159.

19. Miralles F, Ron D, Baiget M, Felez J, Munoz-Canoves P. Differential regulation of urokinase-type plasminogen activator expression by basicfibroblast growth factor and serum in myogen-esis. Requirement of a common mitogen-activated protein kinase pathway. J Biol Chem.1998;273:2052-2058.

20. Blau HM, Pavlath GK, Hardeman EC, et al. Plas-ticity of the differentiated state. Science. 1985;230:758-766.

21. Gredinger E, Gerber AN, Tamir Y, Tapscott SJ,Bengal E. Mitogen-activated protein kinase path-way is involved in the differentiation of musclecells. J Biol Chem. 1998;273:10436-10444.

22. Wu Z, Woodring P, Bhakta K, et al. p38 and extra-cellular signal-regulated kinases regulate themyogenic program at multiple steps. Mol CellBiol. 2000;20:3951-3964.

23. Lijnen HR, Collen D. Regulation and control ofthe fibrinolytic system. In: Festoff BW, ed. SerineProteases and Their Serpin Inhibitors in the Ner-vous System. New York, NY: Plenum Press;1990:9-20.

24. Bell W, Shapiro S, Martinez J, Nossel HL. Theeffects of ancrod, the coagulating enzyme fromthe venom of Malayan pit viper (A. rhodostoma)or prothrombin and fibrinogen metabolism andfibrinopeptide A release in man. J Lab Clin Med.1978:69:592-604.

25. Romer J, Bugge TH, Pyke C, et al. Impairedwound healing in mice with a disrupted plasmino-gen gene. Nat Med. 1996;2:287-292.

26. Kitching AR, Holdsworth SR, Ploplis VA, et al.Plasminogen and plasminogen activators protectagainst renal injury in crescentic glomerulonephri-tis. J Exp Med. 1997;185:963-968.

27. Orimo S, Hiyamatu E, Arahata K, Sugita H. Analy-sis of inflammatory cells and complement C3 in

PLASMIN IS REQUIRED FOR MYOGENESIS 2843BLOOD, 15 APRIL 2002 z VOLUME 99, NUMBER 8

For personal use only. by guest on June 12, 2013. bloodjournal.hematologylibrary.orgFrom

bupivacaine-induced myonecrosis. MuscleNerve. 1991;14:515-520.

28. Gyetko MR, Chen GH, McDonald RA, et al.Urokinase is required for the pulmonary inflam-matory response to Cryptococcus neoformans. Amurine transgenic model. J Clin Invest. 1996;97:1818-1826.

29. Ploplis VA, French EL, Carmeliet P, Collen D,Plow EF. Plasminogen deficiency differentiallyaffects recruitment of inflammatory cell popula-tions in mice. Blood. 1998;91:2005-2009.

30. Sitrin RG, Shollenberger SB, Strieter RM, GyetkoMR. Endogenously produced urokinase amplifiestumor necrosis factor-a secretion by THP-1mononuclear phagocytes. J Leukoc Biol. 1996;59:302-311.

31. Matsushima K, Taguchi M, Kovacs EJ, Young HA,

Oppenheim JJ. Intracellular localization of humanmonocyte associated interleukin 1 (IL1) activityand release of biologically active IL1 from mono-cytes by trypsin and plasmin. J Immunol. 1986;136:2883-2891.

32. Keski-Oja J, Koli K. Enhanced production of plas-minogen activator activity in human and murinekeratinocytes by transforming growth factor-b1.J Invest Dermatol. 1992;99:193-200.

33. Robertson TA, Maley MAL, Grounds MD, Papa-dimitriou JM. The role of macropahges in skeletalmuscle regeneration with particular reference tochemotaxis. Exp Cell Res. 1993;207:321-331.

34. Naldini L, Tamagnone L, Vigna E, et al. Extracel-lular proteolytic cleavage by urokinase is requiredfor activation of hepatocyte growth factor/scatterfactor. EMBO J. 1992;11:4825-4833.

35. Odekon LE, Blasi F, Rifkin DB. Requirement forreceptor-bound urokinase in plasmin-dependentcellular conversion of latent TGF-b to TGF-b.J Cell Physiol. 1994;158:398-407.

36. Tatsumi R, Anderson JE, Nevoret CJ, Halevy O,Allen RE. HGF/SF is present in normal adult skel-etal muscle and is capable of activating satellitecells. Dev Biol. 1998;194:114-128.

37. Shefer G, Oron U, Irintchev A, Halevy O. Skeletalmuscle cell activation by low-energy laser irradia-tion: a role for the MAPK/ERK pathway. J CellPhysiol. 2001;187:73-80.

38. Megeney LA, Kablar B, Garrett K, Anderson JE,Rudnicki MA. MyoD is required for myogenicstem cell function in adult skeletal muscle. GenesDev. 1996;10:1173-1183.

2844 SUELVES et al BLOOD, 15 APRIL 2002 z VOLUME 99, NUMBER 8

For personal use only. by guest on June 12, 2013. bloodjournal.hematologylibrary.orgFrom