Catherine MOALI Extracellular and cell surface proteases ......Wound healing is a complex example of tissue remodelling that includes several steps occurring either concomitantly or

Catherine MOALIDavid J.S. HULMES

Institut de Biologie et Chimie des Protéines,CNRS/Université de Lyon UMR 5086,7 Passage du Vercors, Lyon, France

Reprints: DJS Hulmes<[email protected]>

Extracellular and cell surface proteasesin wound healing: new players are still emerging

Tissue remodelling results from the concerted action of numerousextracellular and cell surface proteases. These act to synchronize thesynthesis and degradation of the extracellular matrix with the controlof cytokine activity and cell signalling in order to create appropriateenvironments for cell proliferation, migration and differentiation.Wound healing is a complex example of tissue remodelling thatincludes several steps occurring either concomitantly or successivelyduring the process of repair: haemostasis, inflammation, angiogenesis,re-epithelialisation, granulation tissue formation, wound contractionand matrix remodelling. The main extracellular and cell surface pro-teases involved in wound healing are serine proteases, especially plas-min, and metalloproteases of the metzincin family (MMPs, ADAM(TS)s, tolloids, meprins, pappalysins) with cysteine proteases playingless prominent roles. Several regulatory proteins and hundreds of sub-strates have been identified for these proteases, either in vitro or invivo. The aim of this review is not to present an exhaustive list ofproteases and related molecules but to give an overview of the proteo-lytic events that are potentially relevant during tissue repair. Newdevelopments aimed at approaching a more integrative view of allthe molecular events involved in tissue remodelling are also discussed.

Key words: extracellular matrix, metzincin, plasmin, skin, tissueremodelling

T issue remodelling, the acquisition of new tissuearchitecture resulting from the combined actionof cells and their extracellular matrix (ECM), is

a feature of normal development and growth, as well aspathological situations and tissue repair. In most tissues,the amount of ECM equals or exceeds the cellular compo-nent and most of the ECM is synthesized by mesenchymalcells (fibroblasts, myofibroblasts and smooth musclecells). Epithelial and endothelial cells also contribute bysynthesizing specialised ECM structures such as basementmembranes, which have both functional and protectiveroles. Exactly how cells orchestrate the organisation ofthe ECM remains largely unclear but extracellular andcell surface proteases have clearly been identified as keyplayers due to their ability to synchronize matrix synthesisand degradation with the activation and deactivation ofgrowth factors and other cytokines.While normal tissue homeostasis depends on a controlledbalance between synthesis and degradation, loss ofhomeostasis in diseases of the skin, cornea or internalorgans can lead either to excessive production (e.g. fibro-sis, hypertrophic scarring) or excessive destruction (e.g.arthritis, emphysema, chronic ulcers) of ECM. Untilrecently, remodelling was regarded almost exclusivelyfrom the destructive point of view with little attentionbeing paid to synthesis. This is explained by the intensiveresearch for new anti-tumoral agents and by the greatexpectations raised by inhibitors of matrix metalloprotei-

nases (MMPs). Even now, the remodelling literature onlyrarely mentions proteases involved in ECM synthesis andone major aspect of this review is to recognize the impor-tance of proteases such as tolloids and ADAMTSs in thematuration of matrix components.In recent years, using knock-out mice and novel toolssuch as activity-based probes and high-throughput proteo-mics, researchers have achieved real progress towardsunderstanding the “protease web” [1] which links pro-teases to other proteases, to their substrates, their inhibi-tors and their various binding partners. However, itremains a challenge to link substrates with observationsat the cellular or tissue level, and in vivo extrapolationof processing events observed in vitro is made difficultby the intrinsic complexity of the extracellular matrix.Substrates can be masked by interaction partners, theycan adopt different conformations, or be physicallyentrapped or otherwise inaccessible to proteases, due tospatial or temporal separation.In contrast to most other post-translational modifications,proteolytic events are irreversible and therefore fine tun-ing of protease activity is required to avoid tissue destruc-tion. For all types of extracellular proteases, inhibitorshave been described, either broad-spectrum, such as α2-macroglobulin [2], or more specific to a protease family,such as TIMPs for MMPs. More recently, mechanisms forenhancement of proteolytic activity have also beendescribed and, interestingly, these usually provide a

Review Eur J Dermatol 2009; 19 (6): 552-64doi:10.1684/ejd.2009.0770

552 EJD, vol. 19, n° 6, November-December 2009

means to direct protease activity to particular substrates.Other regulatory mechanisms following secretion includepropeptide removal, feedback loops involving processedsubstrates and control of extracellular localisation.In this review, we give examples of these various aspectsand present an overview of the properties and functions ofthe extracellular and cell surface proteases (mainly serineproteases and metalloproteases) involved in woundhealing. This well known case of tissue remodelling illus-trates all the different functions with which extracellularand cell surface proteases are known to be involved:matrix protein synthesis and degradation; release of cryp-tic bioactive fragments; regulation of growth factors andother cytokines; and the control of cell adhesion, migra-tion, apoptosis and signalling.

Extracellular and cell surface proteases:old and new players

The functional importance of proteases is demonstrated bythe fact that more than 2% of the genes present in thehuman genome consist of proteases and protease inhibi-tors. These proteases are classified according to theircatalytic mechanism and to the particular amino acid/cofactor responsible for water activation or nucleophilicattack during hydrolysis. Among the 569 human proteases[3], the most abundant are serine and threonine proteases(176 proteins) and metalloproteases (194 proteins) butcysteine and aspartate proteases also play major roles. Ineach class, subfamilies can be identified according to theirlocalization: intracellular, extracellular or associated withthe cell surface. In this review, the focus will be on theextracellular and cell surface associated proteasesinvolved in wound healing. These cluster into threemain groups: serine proteases (mainly proteases from theplasminogen/plasmin system), cysteine proteases andlastly metalloproteases from the metzincin superfamily,which are by far the most abundant (more than 80 mem-bers if MMPs, ADAM(TS)s, tolloids, meprins and pappa-lysins are included). The key importance of plasmin andmetzincins in skin repair was clearly shown by Lund andco-workers [4] who demonstrated that healing was signif-icantly delayed in wounded mice either deficient in plas-minogen or treated with a broad-spectrum metalloproteaseinhibitor (galardin) but was completely blocked whenboth effects were combined.Each protease family could be the subject of an article inits own right; here we present the main features of theseenzymes and refer the reader to the numerous reviewsnow available for more detailed information.

Serine and cysteine proteases

In contrast to the metalloproteases that will be describedbelow, serine proteases do not form a homogeneousfamily with large structural and regulatory differences.Most of the proteases found in the blood circulation(mainly kallikreins [5, 6] and proteases of the coagulationand complement cascades) are serine proteases which canpotentially impact remodelling processes such as angio-genesis and wound healing. Also important are membersof the relatively new family of type II transmembrane

serine proteases (TTSPs; [7]) among which the matrip-tases play an important role in wound healing throughactivation of pro-hepatocyte growth factor/scatter factor[8]. Neutrophil elastase seems to play a major role in theonset of inflammatory responses, especially in the lungwhere, in the absence of tight regulation, changes inneutrophil elastase activity can lead to excessive matrixdegradation (emphysema) or excessive matrix production(lung fibrosis) [9]. Finally, we should mention two serineproteases related to neutrophil elastase, also produced byinflammatory cells, proteinase 3 and cathepsin G [10], aswell as other members of the cathepsin family which arebetter known but belong to the class of the cysteine pro-teases (cathepsins B, K, L, S and V). Cysteine cathepsinscan degrade fibrillar collagens, elastin and some otherextracellular components in acidic pericellular environ-ments [11, 12]. These enzymes have been mainly associ-ated with cardiovascular diseases, cancer progression andbone remodelling; their roles in wound healing remainrelatively elusive.More broadly relevant in the wound healing context are pro-teases from the plasmin system, which are not only respon-sible for fibrin clot lysis [13] but can also participate inTGF-β activation [14, 15], cleave components of the extra-cellular matrix (laminin 332 [16], thrombospondin-1, fibro-nectin [17]) or activate other proteases (especially MMPs)which will in turn degrade ECM. In addition, as mentionedabove, the crucial role of plasmin in tissue repair has beendemonstrated in mice deficient in plasminogen [4, 18].Figure 1 illustrates the complex regulation of plasmino-gen conversion into the active plasmin form [13]. Plasminactivation can be seen as a paradigm for protease regula-tion with both enhancing and inhibiting mechanisms bothcontributing to the fine tuning of protease activity. Theconversion of plasminogen to plasmin is mainly carriedout by two other serine proteases: tPA (tissue-type plas-minogen activator) and uPA (urokinase-type plasminogenactivator) with the latter being the most important for tis-sue remodelling and the former for fibrin clot dissolution.Several modulators of plasmin, tPA and uPA have beendescribed that, in some cases, can act in a substrate-specific manner. Such substrate-specific regulation is fre-quently encountered with proteases, but rarely with othertypes of enzymes. For example, fibrin, the major plasminsubstrate, binds both tPA and plasminogen, therebyincreasing tPA catalytic efficiency and enhancing fibrino-lysis [13]. In addition, uPA can be recruited and activatedmore efficiently at the cell surface via a specific receptoruPAR (uPA receptor). uPAR can then associate with avariety of cell receptors, including integrins, and modulatecell adhesion [19, 20]. Through a positive feedback loop,plasmin cleaves tPA and uPA and the resulting two-chainplasminogen activators are more active than the singlechain proteases. Finally, two inhibitors of the plasminogenactivators (PAI-1 and PAI-2) and inhibitors of plasmin(such as α2-plasmin inhibitor) further regulate plasminactivation. This results in a complex pathway of plasminmaturation, as described in figure 1, accounting for thedestructive potential associated with this protease.

Metalloproteases

Most extracellular metalloproteases belong to themetzincin superfamily which is defined by a conserved

EJD, vol. 19, n° 6, November-December 2009 553

zinc-binding motif (HEXXHXXG/NXXH/D) and amethionine-containing turn near the active site [21, 22].This family is further divided into 6 different subgroups:

astacins which include tolloid proteinases and meprins,matrixins (MMPs), adamalysins (ADAMs andADAMTSs), pappalysins, serralysins and leishmanoly-

Family Substrates/Inhibitors/Enhancers

Serine proteases

FI

FI

HP

E

E

KR Plasminogenactivator (PA)

Plasminogen

+

uPA Receptor(uPAR)

Plasmin

ECM

MMP

ADAMTS

Tolloid (PCP)

Signal peptide Kringle Complement Uegf BMP-1

Variable domain (PLAC, GON-1, PNP, CUB)

Alternative C-terminus

Cleavage site

Transmembrane/cytosolic sequences

GPI anchor

TM/Cs

GPI

Haemopexin-like

Fibronectin type 2

Disintegrin

Cysteine rich

Thrombospondin repeat

Propeptide

Catalytic domain

Epidermal growth factor

Fibronectin type I

PCPEnhancer(PCPE)

Twistedgastrulation

Olfactomedin

growth factorsECM

+ +

Fibulin-1Tissue Inhibitorof MMP (TIMP)

Papilin

ECM

+

Tissue Inhibitorof MMP (TIMP)

growth factors

adhesion molecules

ECM

receptors

MMPs

MMPs

TGF-β

Fibrin(ogen)

PA inhibitor(PAI)

α2-Plasmininhibitor

KR

KR

KR

Pro

Pro

Pro

Zn

Zn DI CR

E

E

E

F2

E

Pro

Pro

Pro

Zn

Zn CUB CUB CUB CUB CUB

CUB

DI

DI

TS CR

CR

n

TM/Cs

TS TS

TS

TS

TM/Csor GPI

Zn

(F2)3

(HP)4

(HP)4

Pro KR KR KR

KR

KR

S tPA

uPA

Plasminogen

MMPs

MT-MMPs

ADAMs

ADAMTSs

Tolloids

Metalloproteases

S

S

Figure 1. Structural features, substrates and possible modulators of extracellular and transmembrane proteases involved inwound repair. Protease inhibitors are shown in orange and enhancers in green.


sins. With the exception of the last two groups, all thesefamilies are involved in remodelling processes.

Matrix metalloproteases (MMPs)

MMPs are, without contest, the most well-known of allremodelling enzymes and have been extensively studiedin several contexts (cancer, cardiovascular diseases,wound healing, bone remodelling, pregnancy…) for thepast 30 years. MMP activity is inherently involved inremodelling since, in normal situations, these enzymesare expressed at virtually undetectable levels with a fewexceptions such as MMP-7 which is constitutivelyexpressed in epithelial cells [23, 24]. Their expressionbecomes elevated when there is a challenge to the system,such as wound healing and disease [25].The initial classification of the 23 human MMPsaccording to their substrate specificity and functionalcharacteristics (collagenases, gelatinases, broad-specificity stromelysins, membrane-bound MMPs…) hasnow given way to a simple numbering scheme. While allMMPs have in common a signal peptide, a propeptide anda catalytic domain, they differ by the possible insertion offibronectin-type II repeats in the catalytic domain (MMPs-2 and -9), the presence of a haemopexin domain involvedin substrate recognition (in all MMPs except MMPs-7 and-26), an additional transmembrane domain followed by acytoplasmic tail (MMPs-14, -15, -16 and 24) or a GPIanchor (MMPs-17 and -25) [26]. The propeptide iscleaved off either in the Golgi apparatus by proteasesfrom the furin-like proprotein convertase family or peri-cellularly by plasmin, other MMPs or by autoproteolysis.As a family, MMPs can cleave all extracellular matrixcomponents (collagens, elastin, proteoglycans and otherstructural glycoproteins). They are also involved ingrowth factor activation (e.g. TGF-β and VEGF: seebelow) and can modify cell-matrix and cell-cell interac-tions in several ways (including matrix degradation, shed-ding of syndecans and cadherins, and binding of integrinsand CD44) [27, 28]. After propeptide removal, MMPactivity is regulated by potent inhibitors of the TIMP fam-ily (Tissue Inhibitors of MMPs 1-4) which have also beenimplicated in a range of biological processes, grosslyoverlapping MMP functions but with opposite effects:tumour metastasis, angiogenesis, wound healing, cell dif-ferentiation and survival, inflammatory responses, etc.TIMPs bind to the catalytic zinc ion of MMPs, thusblocking access to the catalytic pocket. All MMPs canbe inhibited by all TIMPs but to variable extents [29].The main structural and regulatory features of this familyare also summarized in figure 1.

ADAMs and ADAMTSs

The adamalysin family is the largest family of metallopro-teases in terms of numbers with 40 human ADAMs (ADisintegrin And Metalloprotease) and 19 ADAMTSs(ADAMs with thrombospondin (TS) motifs). Among theADAMs, only 13 are predicted to be catalytically active,and many have lost the ability to bind zinc, pointing toadditional, non-enzymatic roles, for example in cell adhe-sion. ADAMs and ADAMTSs share some structural sim-ilarities (a catalytic domain followed by a disintegrindomain and a cysteine-rich domain: figure 1) but majordifferences in their cellular localizations and substrate spe-cificities have led to divergent biological functions.

With a few exceptions, ADAM proteins remain attachedto the plasma membrane and the catalytic activity of theADAMs that are predicted to be active (8-10, 12, 15, 17,19-21, 28, 30, 33) is devoted to the shedding of the ecto-domain of several transmembrane proteins. They releasediffusible factors, such as heparin-binding EGF (HB-EGF) or TNF-α, from their membrane-anchored precur-sors. They also cleave adhesion molecules (e.g. cadherinsand CD44) to disrupt adhesive contacts while shedding ofsignalling receptors (e.g. the Notch receptor) leads to fur-ther intra-membrane proteolysis and release of an intracel-lular domain which becomes a signal transducer (seebelow). In the latter case, binding of a substrate ligand,such as Delta in the case of the Notch receptor, is oftenrequired to trigger proteolysis by ADAMs and permitsvery strict regulation of the downstream signalling path-way. The ADAMs that are proteolytically inactive cannevertheless play important roles, especially in cell adhe-sion, via their disintegrin and cysteine-rich domains. Thedisintegrin domains of several ADAMs, for example,have been shown to bind to receptors of the integrin fam-ily and thereby block their function. Furthermore, thecysteine-rich domain can interact with ECM proteinssuch as fibronectin and heparan-sulphate proteoglycansof the syndecan family.In contrast to the ADAMs, ADAMTSs are all secretedproteases and are found either in soluble form or boundto the ECM. Their key feature is a variable number oftype 1 thrombospondin repeats (figure 1). They are allsupposedly active but six of the ADAMTS are orphanswith no known substrates. Other ADAMTSs tend to bemuch more specific than the metalloproteases describedso far and their main remodelling functions can be classi-fied into 3 overlapping groups [30]: (i) the anti-angiogenicADAMTSs (-1 and -8), this function being directly medi-ated by their TS motifs or through the release of anti-angiogenic factors such as thrombospondin fragments[31] (ii) the aggrecanases (-1, -4, -5, -8, -9 and -15)which cleave aggrecan and are the main players duringjoint degeneration and (iii) the procollagen N-proteinases(-2, -3 and -14) which are involved in processing of fibril-lar collagen precursors.For both ADAMs and ADAMTSs, propeptide removal iscarried out by proteases of the proprotein convertasefamily in the Golgi apparatus. The only protein inhibitorsthat have been described so far for ADAMs are the TIMPsbut in contrast to the MMPs, not all members of the fam-ily are inhibited by the four TIMPs and the physiologicalconsequences of this inhibition are not very well docu-mented. TIMP-3, which is unique among the TIMPs forits ability to interact strongly with ECM, is also a potentinhibitor of the catalytic domain of ADAM-10 and -17[32, 33]. In the ADAMTS family, full-length ADAMTS-4 and -5 have also been shown to be inhibited by TIMP-3[34]. Interestingly, TIMP-3 is not a good inhibitor ofADAMTS-2 but this particular enzyme has been shownto be inhibited by papilin [35], a protein found in thebasement membrane which contains the C-terminal, noncatalytic domains of ADAMTSs. Finally, ADAMTS-1-dependent cleavage of aggrecan seems to be enhancedby fibulin-1 through simultaneous binding of both prote-ase and substrate [36].For a more comprehensive overview of the members androles of both families, we recommend the recent reviews


by Tousseyn et al. and Edwards et al. on ADAMs [37, 38]and Porter et al. on ADAMTSs [30].

Tolloids

Tolloid proteinases are also known as procollagen C-proteinases (PCPs) or by the names of the four membersof the family (BMP-1, mTLD, mTLL-1 and mTLL-2).BMP-1 (bone morphogenetic protein-1) and mTLD(mammalian tolloid) are splice variants of a single genewhereas mTLL-1 and 2 (mammalian tolloid like-1 and -2)are the products of two separate genes. Members of thisfamily were originally linked exclusively to collagenmaturation as one of their major roles is to cleave theC-terminal propeptides of procollagens I, II and III,thereby triggering fibril formation. However, a variety ofnew substrates have been discovered in the past decade[39] and tolloids now appear as key players in variousbiological functions ranging from embryogenesis andmorphogenesis to ECM biosynthesis. For example, theycontrol dorso-ventral patterning in the embryo throughthe cleavage of chordin, a growth factor antagonist, aswell as morphogenetic events such as muscle growthand neural differentiation through the activation of GDF-8 and -11. It was also recently established that tolloids cancleave latent TGF-β binding proteins (LTBPs) therebyreleasing latent TGF-β from the matrix and triggeringTGF-β activation [40] (figure 2). In addition, they process

several partners involved in collagen fibrillogenesis andparticipate in basement membrane assembly through theprocessing of laminin 332 [41] and procollagen VII [42].Finally, tolloid proteinases probably also play a role inangiogenesis with the recent discovery of substrates suchas endorepellin [43] and prolactin [44]. A unified viewsuggesting that this family of proteases could act to syn-chronize signalling pathways with ECM biosynthesis, acrucial requirement during morphogenesis and tissueremodelling, is now emerging.From a structural point of view, tolloid proteinases com-prise a catalytic domain of the astacin type and a variablenumber of CUB and EGF domains (figure 1). Their pro-peptides are also removed by furin-like proteases, as forsome MMPs and almost all ADAM(TS)s. Tolloids aresoluble proteins which are regulated substrate-specificenhancers. In addition, in Xenopus, it has been shownthat the Sizzled protein binds strongly to tolloids and inhi-bits cleavage of chordin [45], suggesting that homologoussecreted frizzled-related proteins might play similar rolesin mammals (though not supported by recent data, [46]).Tolloids are also regulated by enhancer proteins which aresubstrate-specific in that they specifically activate thecleavage of only one type of substrate, for example pro-collagen C-proteinase enhancers (PCPEs) in the case ofthe fibrillar procollagen substrates [47] or twisted gastru-lation and olfactomedin in the case of chordin [48, 49].

Procollagens I, III (V)

Elastin

Laminin 332

LAP

ActiveTGF-β

ECM

LTBP

TGF-β

PerlecanSmall leucine richproteoglycans

Endorepellin

Prolysyloxidase

Prolactin, growth hormone

Tolloids

Tolloids

Tolloids

Tolloids

Tolloids

TolloidsPlasmin

MMP 2, 9, 14Tolloids

Tolloids

Procollagen VII

Inhibition of angiogenesis

Basement membrane assembly and control of cell migrationCollagen and elastin fibrillogenesis

TGF-β activation

TolloidsUnknown

Tolloids(furin)

ADAMTS 2, 3, 14(tolloids)

γ2β3

α3

Figure 2. Roles of tolloid proteinases in wound healing. Tolloid proteinases act on several substrates, either alone or in com-bination with other types of protease (cleavage sites indicated by scissors), in order to regulate collagen and elastin fibrillo-genesis, angiogenesis, basement membrane assembly, cell migration and growth factor activation. ECM = extracellular matrix;LTBP = latent transforming growth factor binding protein; LAP = latency associated peptide.


Other metalloproteases

Two other families of metzincins are also slowly makingtheir way into the remodelling field: pappalysins andmeprins. The best studied member of the pappalysin fam-ily is PAPP-A (Pregnancy-Associated Plasma Protein-A)which cleaves IGF binding proteins 2, 4 and 5, therebyreleasing sequestered IGF-I and II with effects on cellproliferation, migration and differentiation. Elevatedlevels of this protease are found in the circulation of preg-nant women and of patients with acute coronary syn-dromes. PAPP-A also seems to play a role during woundhealing with an increased expression in the dermis andepidermis [50]. The proteolytic activity of PAPP-A isinhibited, in an unusual and probably irreversible manner,by the proform of eosinophil major basic protein (pro-MBP), which forms a covalent 2:2 proteinase-inhibitorcomplex based on disulfide bonds [51]. Interestinglyalso, upon IGF binding, cleavage of IGFBP-2 and 4 isenhanced many fold while cleavage of IGFBP-5 isslightly inhibited [52].Meprins, like tolloid enzymes, are proteases of the astacinfamily [21], which, unlike tolloid proteases, are synthe-sised as transmembrane proteins. They are expressed astwo independent subunits (α and β) which can assembleinto hetero- or homo-dimers, or higher oligomers in thecase of meprin α, which can be released from the plasmamembrane by proteolytic shedding. Expression of meprinshas been reported mainly in the kidney, intestine and skinbut could increase during inflammatory conditions, cancerprogression and wound healing. In recent years, meprinshave been shown to cleave several ECM proteins (colla-gen IV, laminin 111, nidogen-1, fibronectin), enzymessuch as lysyl oxidases and adhesion proteins such as cad-herins but the discovery of physiologically relevant sub-strates is still a matter of intense research. The onlyendogenous inhibitor described so far for meprins ismannan-binding lectin [53] but the primary regulatorymechanism of meprins after propeptide removal seemsto be targeting to specific locations (e.g. apical membraneof polarized cells).

Extracellular and cell surface proteasesin wound healing

The most common case of tissue remodelling, tissuerepair, occurs after various types of traumatic lesions(wounds, burns, surgery…). Tissue repair is usuallydescribed as being divided into three phases which par-tially overlap, while the entire process can take severalmonths.During the first phase, haemostasis and inflammation, afibrin-rich provisional matrix is formed as a result ofthrombin and platelet activation. Fragments of coagulationfactors, activated complement components, growth factorsand chemokines released by damaged cells and activatedplatelets then recruit inflammatory cells to the site ofinjury [54, 55]. Platelets, for example, are major sourcesof TGF-β and PDGF. First to arrive are the neutrophils,which cleanse the wound area by phagocytosis and therelease of antimicrobial substances (reactive oxygen spe-cies, cationic peptides, eicosanoids) and proteases (neutro-phil elastase, cathepsin G, proteinase 3, uPA, MMPs-8

and -9) [56]. Migration of neutrophils from the bloodcapillaries into the wound site involves interactions withendothelial cell adhesion molecules such as selectins andICAMs. Subseqently, monocytes also infiltrate the woundsite, in response to growth factors (TGF-β, PDGF) andchemokines (e.g. monocyte chemoattractant proteins orMCPs) where they become activated and differentiateinto mature tissue macrophages. While both neutrophilsand macrophages produce pro-inflammatory mediators(TNF-α, interleukins-1 and -6), macrophages also synthe-size several growth factors (TGF-β, TGF-α, bFGF, PDGF,VEGF) to stimulate the repair response.In the second phase, re-epithelialisation and granulationtissue formation, which commences soon after injury,keratinocytes use the provisional fibrin-rich matrix tomigrate and cover the wound area. This requires the coor-dinated action of several growth factors and cytokines,together with multiple proteases and adhesion proteinsinvolved in cell-cell and cell-matrix interactions. Impor-tant growth factors involved in re-epithelialisation includemembers of the EGF family (EGF, TGF-α, HB-EGF) andFGFs, which promote keratinocyte proliferation, andTGF-β, which negatively regulates this process [57].Next, about 4 days after injury, fibroblasts are alsoattracted to the wound area to initiate the formation ofgranulation tissue, which includes a new extracellularmatrix rich in collagen III, fibronectin and proteoglycans.Cross-talk between keratinocytes and fibroblasts plays animportant role, notably in the production of a new base-ment membrane, as well as in the control of cell prolifer-ation. Fibroblast proliferation is stimulated by differentgrowth factors (including PDGF, TGF-β, CTGF, IGFs)some of which also promote the important process ofwound contraction by stimulating fibroblast differentia-tion into myofibroblasts. In addition to the acquisition ofa contractile phenotype, these myofibroblasts also demon-strate an enhanced ability to synthesize extracellularmatrix. Finally, formation of granulation tissue alsorequires new blood vessel formation, angiogenesis,under the control of different growth factors (notablyVEGF) as well as anti-angiogenic factors released by pro-teolysis (see below).In the final phase of tissue repair, tissue remodelling, thegranulation tissue is progressively replaced by a moremature scar tissue, where collagen I is the major compo-nent. Remodelling is a lengthy process, which requiresboth synthesis and degradation of ECM in order torecreate as closely as possible, but never completely, theproperties of the original tissue. At the end of woundrepair, myofibroblasts usually undergo apoptosis and arereplaced by normal fibroblasts. However, in certainpathologies, excessive inflammation can lead to persis-tence of myofibroblasts and excessive ECM synthesis,leading to hypertrophic scarring, or complete failure toheal, as in chronic wounds (ulcers).Proteases play numerous roles during the process ofwound healing (table 1). They are at work during: hae-mostasis; regulation of the activities of growth factors,chemokines and other cytokines; cell migration, signallingand survival; angiogenesis; and synthesis and remodellingof the extracellular matrix.


Haemostasis and inflammation

Upon injury, damage to the blood vessel wall exposessub-endothelial proteins, notably collagens, to which pla-telets bind via specific glycoprotein receptors, resulting inplatelet activation and subsequent aggregation. Formationof the platelet plug is stabilised by interaction with vonWillebrand factor, which itself is a substrate forADAMTS-13 [30]. In addition, negative regulation ofaggregation occurs by proteolytic release from the cellsurface, or shedding, of platelet glycoprotein receptors byADAM-17/TACE [58]. The platelet plug is further stabi-lised by the formation of a fibrin-rich network resultingfrom thrombin activation, via the coagulation cascade,and subsequent maturation of fibrinogen.Circulating leukocytes then infiltrate into the wound areafirst by attachment to the endothelial cells of the bloodvessel wall followed by extravasation into the surroundingtissue. This involves tethering to specific cell adhesionmolecules expressed on the endothelial cells, notably E-selectin, ICAM-1 and VCAM-1, whose expression isinduced by inflammatory cytokines [59]. Subsequentextravasation through the blood vessel wall requires addi-tional cell adhesion molecules such as PECAM-1 [60].These processes are under the control of multiple proteases

(notably ADAMs-10 and -17) through shedding of ICAM-1, VCAM-1 and PECAM-1 from endothelial cells, as wellas by shedding of L-selectin from neutrophils. Finally,neutrophil infiltration into the wound site is facilitated bysecretion of the neutrophil collagenase MMP-8, probablythrough collagen degradation but also through proteolyticmodification of chemotactic molecules [61, 62].Migration of inflammatory cells into the wound site isactually controlled by several chemokines, many ofwhich are regulated by the action of proteases [63, 64].Protease activity can lead to increased or decreasedactivity, either directly, by proteolytic cleavage of the che-mokine, or indirectly, by cleavage of interaction partnersor receptors. Particularly well studied are the monocytechemoattractant proteins (MCPs) which serve to attractmonocytes to the site of tissue injury. A number ofMMPs have been found to cleave short N-terminal frag-ments from different MCPs leading to a loss of chemokineactivity and the generation of receptor antagonists [65].Since these MMPs are produced by stromal cells as aresult of macrophage-derived inflammatory cytokines,such activity could serve to dampen the inflammatoryresponse. Alternatively, other chemokines can be directlyactivated by proteolytic activity, either by the processing

Table 1. Biological functions of extracellular and cell surface proteases in tissue remodelling: substrates and proteasesinvolved. Compiled with the help of the CutDB proteolytic event database (http://cutdb.burnham.org/), from where originalreferences can be obtained

Function Substrates Proteases involved

Matrix assembly procollagens, prolysyl oxidases, osteoglycin, biglycan tolloids, ADAMTSsMatrix degradation fibrillar collagens MMPs

fibronectin, vitronectin MMPs, kallikreins, plasmin, meprins, ADAMsCOMP, matrilin-3 ADAMTSstenascins, SPARC/osteonectin, osteopontin, thrombospondin ADAMTSs, MMPs, plasmin, meprins,

cathepsin Gaggrecan, brevican, versican MMPs, ADAMTSs, plasmindecorin, fibromodulin, lumican, biglycan, osteoglycin MMPs, ADAMTSselastin, fibrillins, fibulin MMPsfibrinogen plasmin, MMPs, kallikreinslaminins, nidogens, collagen IV, collagen XVIII MMPs, kallikreins, cathepsin S, meprins,

ADAMs, plasminProtease activation/inhibition

ProMMPs, plasminogen, meprins MMPs, uPA, tPA, plasmin, kallikreins,autocatalysis

serpins, elafin, TIMP-1 MMPs, kallikreins, cathepsins, neutrophilelastase

PCPE-1 MMP-2Release of bioactivefragments

collagen IV (tumstatin), collagens XV/XVIII (endostatins),perlecan (endorepellin), plasmin (angiostatin), prolactin,growth hormone, thrombospondin

MMPs, plasmin, cathepsins, tolloids,ADAMTS-1

Control of cell adhesion/migration

laminins, cadherin, L-selectin, ICAM-1, VCAM-1,PECAM-1, L1, CD44

MMPs, tolloids, plasmin, neutrophil elastase,ADAMs, cathepsin G

Release/activation/inactivation of growthfactors, chemokines andcytokines

TGF-β, TGF-α, TNF-α, IL-8 (CXCL8), IL-1β, CTGF, PDGF,VEGF, monocyte chemoattractant proteins (CCLs-2,-7,-8,-13),HB-EGF, GDF8, GDF11, IGFBPs, chordin, LTBP, CX3CL1;CCLs-5,-20,21; CXCLs-1,-4,-5,-6,-9,-10,-16

MMPs, tolloids, plasmin, ADAMs, cathepsins,neutrophil elastase, uPA

Shedding of cell surfacereceptors and otherproteins

Estrogen receptor, TNF receptor, IL-6/IL-15 receptors,macrophage colony stimulating factor receptor, NGF receptor,growth hormone receptor, EGF receptor, FGF receptor,integrins, CXCR4, CD44, LRP8, platelet glycoproteins,syndecans, transglutaminase 2, CD44, collagen XVII

ADAMs, MMPs, neutrophil elastase,cathepsins

Cell survival Fas-ligand, stromal cell derived factor-1 (CXCL12) MMPs, cathespin G, neutrophil elastaseCell signalling Notch/delta, protease activated receptors ADAMs, thrombin


of inactive precursors (such as IL-8/CXCL8 activation byserine proteases, cathepsin L or MMP-9) or by sheddingof membrane-bound precursors (such as CX3CL1/frac-talkine by MMP-2 [66], ADAMs-10 and -17 [67]). Fur-thermore, since IL-8 binds to the heparan sulphate chainsof cell surface syndecan-1, MMP-7 shedding of the IL-8/syndecan-1 complex can help establish local chemokinegradients at the wound site [63].Finally, ADAM-17/TACE also plays a key role in theregulation of proinflammatory cytokines, both by theshedding of TNF-α from the cell surface, as well as bythe shedding of both TNF-α and IL-6 receptors [38]. Sim-ilarly, MMPs-7 and -12 may also release TNF-α frommacrophages, while MMPs-2,-3 and -9 are involved inthe activation and inactivation of IL-1β [63].

Re-epithelialisation, angiogenesis and granulation tissueformation

Re-epithelialisation, angiogenesis and granulation tissueformation are also governed by the orchestrated action ofnumerous growth factors as well as multiple proteasesinvolved in growth factor release from the matrix orfrom the cell surface, control of cell signalling and modu-lation of cell-matrix and cell-cell interactions. Differentgrowth factors are involved in each step, including TGF-α, HB-EGF and KGF in epithelial cell (keratinocyte) pro-liferation and migration; bFGF, VEGF and TGF-β inangiogenesis; and TGF-β, PDGF, CTGF and IGF in fibro-blast proliferation and ECM production; while othergrowth factors, such as HGF/SF, have more general effects[8, 57].Many of the examples described below show the impor-tance of proteolytic processing at the cell surface, not onlyfor transmembrane proteases (ADAMs, meprins, mem-brane type MMPs, matriptases) but also for other pro-teases that are anchored to the plasma membrane, eitherdirectly, through a GPI anchor (MMPs-17,-25), or indi-rectly, in association with cell surface receptors. Examplesof the latter are binding of MMP-9 to the hyaluronanreceptor CD44 [68], MMP-2 to integrin αvβ3, MMP-1 tointegrin α2β1 and MMP-7 to heparan sulphate proteogly-cans [63]. Such interactions serve to localise the site ofproteolysis, or protease activation, as well as to bring pro-teases and their substrates in closer proximity.

Release of growth factors sequestered in the extracellular matrix

One function of proteases is to mobilise growth factors,such as VEGF, FGF, PDGF, CTGF and TGF-β, by releasefrom matrix-bound stores [27, 69, 70]. For example,angiogenic factors such as VEGF and bFGF associatewith heparan sulphate proteoglycans found in the matrixand must be released to promote angiogenesis. MMP-9 isthought to be the main enzyme involved in the release ofVEGF [71] but other mechanisms involving MMP-2 havebeen described [25, 72]. Interestingly, the released andmatrix-bound forms of VEGF have different properties,the former inducing endothelial cell proliferation, the lat-ter stimulating vascular sprouting and branching [73].MMP-2 can also unmask VEGF by cleaving inhibitorypartners such as pleiotrophin/HARP and CTGF [74].Similarly, IGFs are antagonized by six different IGF bind-ing proteins that can be cleaved by pappalysins, MMPs,cathepsins, kallikreins or other serine proteases [75].

The activation of TGF-β has been extensively studied, andseveral mechanisms, both proteolytic and non-proteolytic,have been elucidated. TGF-β, (of which there are threevariants β1, β2 and β3), is secreted as a latent dimericcomplex consisting of the functional growth factor at theC-terminus and the latent prodomain (called LAP orlatency-associated peptide) at the N-terminus. TGF- β iscleaved from its propeptide during secretion by furin-likeproteases and remains bound to LAP in a non-covalentmanner. LAP can also form disulphide bridges with latentTGF- β binding proteins LTBPs-1, -3 or -4 [76] to pro-duce LLC (large latent complex). Because LTBPs have astrong affinity for several matrix components, especiallyfibrillin-1 [77, 78], LLC is deposited within the extracel-lular matrix, to which it is covalently bound by transglu-taminase cross-linking [79]. Release of the TGF-β/LAPcomplex from the ECM requires proteolysis of the LTBPand this processing was recently attributed to tolloids [40](Fig.2). Subsequently, active TGF-β can be released fromthe complex by another round of proteolysis, this timeunder the control of plasmin [14] or MMPs-2, -9 and -14, themselves possibly in association with integrins, towhich LAP also binds [68, 79, 80]. Conformationalchanges leading to TGF-β activation in the absence ofproteolytic activity have also been demonstrated, thefirst being mediated by thrombospondin-1 (TSP-1)which can bind to LAP associated with TGF-β [81].This interaction appears to induce a conformationalchange in the prodomain which, while remaining boundto mature TGF-β, leads to a loss of latency and henceTGF-β activation [82]. More recently, it has been shownthat mechanical forces exerted by myofibroblasts attachedto an extracellular matrix substrate can directly lead toTGF-β activation, through an LTBP-dependent but non-proteolytic pathway [83]. With LAP bound to differenttypes of integrins via its RGD motif, cell shape changeor movement can distort the structure of the latent com-plex, thereby releasing active TGF-β [79]. Interestingly,this mechanism only works when cells remain bound toa relatively rigid substrate, in order to provide sufficientmechanical traction, thereby demonstrating the impor-tance of the physical properties of the extracellular matrixin growth factor activation.

Release of membrane-bound growth factors and control of cellsignalling

Growth factors can also be activated by proteolyticrelease, or shedding, from the cell surface. For example,growth factors of the EGF family, notably TGF-α andheparin binding (HB)-EGF have been shown to playimportant roles in wound healing, particularly in re-epithelialisation [57]. All EGF receptor ligands are syn-thesized as membrane-anchored forms, which are acti-vated by ectodomain shedding from the cell surface.One of the best described examples is the shedding ofHB-EGF by ADAMs-9, -10, -12 or -17 [84]. HB-EGF isexpressed on the cell surface of keratinocytes and in theevent of injury, expression of ADAMs leads to sheddingof the growth factor and transient stimulation of the EGFreceptor which is required for keratinocyte migration. Inaddition, shedding of the extracellular domain of HB-EGFalso results in the release of a C-terminal cytosolic frag-ment that can be translocated to the nucleus where it


associates with transcription factors involved in cell cyclecontrol [85].In addition to growth factor activation, shedding alsoplays a role in cell signalling. One of the best studiedexamples is the Notch signalling pathway which isinvolved in a large panel of developmental events and inseveral cancers [86] and was recently shown to be directlyinvolved in wound healing [87]. Thus, when Notch sig-nalling is blocked, delayed healing is observed, due tochanges in endothelial cell, keratinocyte and fibroblastmigration. Both the Notch receptor and its ligands (mem-bers of the Delta and Jagged families) are transmembraneproteins with large extracellular domains and as a conse-quence, interaction between receptor and ligands isrestricted to neighbouring cells. When a ligand is boundto the Notch receptor, two sequential proteolytic eventsare induced: the first cleavage is performed by ADAM-10 or -17 resulting in the release of the Notch extracellulardomain, while the second cleavage occurs inside themembrane, by regulated intracellular proteolysis or RIP-ping, and is due to the γ-secretase complex [88]. The sec-ond cleavage promotes the translocation of the Notchintracellular domain into the nucleus with subsequentinduction of transcriptional events.Another example of the importance of proteases in cellsignalling has been described for the family of proteaseactivated receptors (PARs). These are transmembrane G-protein coupled receptors in which cleavage of part of theextracellular domain (usually by serine proteases) exposesa tethered ligand (part of the polypeptide chain) whichthen interacts with the rest of the molecule, thereby trig-gering intracellular signalling in an irreversible manner.Activation of PARs by thrombin can have a direct effecton the expression of profibrotic mediators such as PDGFand CTGF [89], as well as TGF-β [90] and has beenshown to be involved in tissue fibrosis.

Modulation of cell-matrix and cell-cell interactions

Both re-epithelialisation and angiogenesis involve severalproteolytically regulated steps, including cell migrationthrough the provisional matrix, and disruption/re-establishment of both cell-cell and cell-matrix contacts.For example, epithelial migration through the fibrin-richmatrix is in part due to activation of serum-derived plas-min by up-regulation of uPA and its receptor uPAR.Plasminogen-deficient mice show severely impaired re-epithelialization in a skin wound model [91]. Similarly,laminin 332 (previously known as laminin 5) which con-nects basement membranes to integrins and syndecans onthe epithelial cell surface can be cleaved by several pro-teases, including serine proteases, MMPs and tolloid pro-teases. These proteolytic events generally result inincreased or decreased migration of epithelial cells,depending on which chain is processed and at which site[92, 93].With regard to cell-cell interactions, those mediated bycadherins (E-cadherin in the case of keratinocytes or VE-cadherin the case of endothelial cells) are also subject toprotease regulated shedding. Several proteases are poten-tially involved, including MMPs-3,-7,-9,-12 and-14 [94,95], ADAM-10 [96], neutrophil elastase [97] and meprin-β [98]. Shedding of cadherins can also result in transloca-tion of membrane bound β-catenin to the cell nucleus,thereby leading to changes in cell proliferation [99].

Another effect of ECM proteolysis is to expose crypticsites or fragments (matricryptins) that change cell-matrixinteractions [100, 101]. For example, PDGF-stimulatedvascular smooth muscle cells migrate more rapidly on col-lagen I fragments generated by collagenase activity thanon intact collagen [102]. This is due to changes in adhe-sion and migration mediated by a switch in integrinrecognition from α2β1 to αvβ3. Similarly, MMP cleavageof basement membrane collagen IV exposes a cryptic siteassociated with increased angiogenesis coupled withdecreased binding to integrin α1β1 and increased bindingto αvβ3 [103]. Further examples are the release of anti-angiogenic fragments from pro-angiogenic molecules oreven from proteases themselves, including: (i) angiostatinfrom plasmin [104], (ii) anti-angiogenic fragments frombasement membrane components, including endostatinfrom collagen XVIII, tumstatin, canstatin and arrestenfrom collagen IV and endorepellin from perlecan [43,101], and (iii) anti-angiogenic fragments from prolactinand growth hormone [44] (figure 2).Finally, in addition to their effects on cell proliferation,signalling, adhesion and migration, extracellular proteasesare also known to influence programmed cell death, apo-ptosis [105]. Proteolytic disruption of cell-matrix interac-tions can lead to apoptosis by the process of anoikis. Alter-natively, apoptosis can be induced by shedding of Fasligand, or inhibited by release of its “death” receptor Fas/CD95, both of which have been reported for MMP-7 [24].

Matrix remodelling

In general, relatively little is known, in structural terms,about how cells remodel their surrounding ECM. Clearlyboth mechanical as well as biochemical mechanisms areinvolved, as shown for example by the different signallingmechanisms involved in contracting floating andrestrained collagen gels [106], or by the inability ofTGF-β to induce myofibroblast differentiation unless tis-sues are under mechanical stress [107]. The mechanicalproperties of tissues such as skin are in turn largely deter-mined by the composition and organization of the ECM,itself regulated by proteolytic and cross-linking enzymes.Matrix stiffness is a key factor in cell migration; thematrix should be stiff enough to support cell traction butnot too stiff to avoid excess adhesion [108]. Cell migra-tion is usually, but not always, accompanied by proteo-lytic activity, but interestingly recent observations showthat the site of proteolysis is behind the leading edge ofthe cell, thereby allowing cells to attach to the matrixbefore it is degraded [109]. In this way, for example,ECM rigidity promotes increased matrix-degradingactivity of invasive cells at specialized actin-rich protru-sions called invadopodia or podosomes [110].Granulation tissue formation involves the relatively rapiddeposition of fibrillar collagens (mainly collagen III),fibronectin, proteoglycans and other proteins in a newlysynthesized ECM. The tolloids are the group of extracel-lular proteases that have been most clearly linked to theprocess of ECM assembly [39] (figure 2). Fibrillar col-lagens (types I, II, III, V and XI) are synthesized in pre-cursor form, with each molecule having both N- and C-terminal propeptide extensions [111, 112]. The propep-tides help maintain solubility during synthesis of the pro-collagen molecule and are removed either just before or


shortly after secretion into the ECM. In the case of themajor fibrillagen collagens (types I, II and III), the C-propeptides are removed by tolloid proteases, while theN-propeptides are removed by members of the ADAMTSfamily (types -2, -3 and -14). Following cleavage of pro-peptides, mature collagen molecules can then spontane-ously assemble into fibrils. Proteolytic processing of pro-collagens V and XI is more complicated, with tolloids alsobeing involved in the cleavage of N-propeptides andfurin-like enzymes contributing to C-terminal maturation[113]. Most collagen fibrils are usually heterotypic innature, i.e. composed of different collagen types, forexamples collagen I/III/V fibrils in skin or collagen II/XIfibrils in cartilage, and it is known that such heterotypicinteractions play a role in the regulation of collagen fibrildiameter [111]. In addition, variations in the rates andextent of processing between different collagen types, par-ticularly in the N-terminal regions, also contribute todiameter regulation. Collagen fibril assembly is furtherregulated by interactions with members of the small leu-cine rich proteoglycans (SLRPs, including decorin, fibro-modulin, biglycan, osteoglycin), among which biglycanand osteoglycin have been shown to be cleaved by tolloidproteases [114, 115]. In the case of osteoglycin, thetolloid-dependent processing potentiates inhibition of col-lagen fibril assembly. Once assembled, collagen fibrils arestabilised by the formation of covalent crosslinks betweenneighbouring molecules, in a process catalysed by cross-linking enzymes such as lysyl oxidases and transglutami-nases [116, 117]. Lysyl oxidases are themselves synthe-sized in a precursor form, and can be activated by tolloidprotease cleavage of propeptides [118]. In addition, lysyloxidases also act on elastin fibres as well as in bFGF andTGF-β signalling [119]. With regard to the basementmembrane, tolloid proteases regulate the assembly ofanchoring filaments, involved in stabilizing the dermal-epidermal junction, by propeptide cleavage of the non-fibrillar procollagen VII [42], as well as incorporation oflaminin 332, by processing of the γ2 chain [41] (figure 2).Bearing in mind that tolloid proteases also catalyze thefirst step of TGF-β release from the matrix and thatTGF-β increases the expression of most matrix compo-nents, tolloid proteases appear therefore as major hubs inthe control of ECM assembly.Conversion of granulation tissue into the mature form is aprocess that can take several months and rarely results inthe same structural and functional properties as those ofthe original tissue [54]. Remodelling requires the coordi-nated action of numerous proteases involved in ECMdegradation and synthesis (essentially MMPs andADAMTSs for the degradative part; table 1) as well asin the control of cell proliferation, differentiation, adhe-sion, migration and death, through the various mechan-isms described above. In adults, the end result of theremodelling phase is often a scar [120] which reflects dif-ferences in the processes of repair and regeneration, wherethe former has evolved as a means to escape danger byrapid recovery of tissue integrity [121]. Scars are charac-terized by dense bundles of mostly type I collagen fibrilswhich appear to form during the process of wound con-traction and become oriented along lines of mechanicalstress [122, 123]. Wound contraction is thought to bedue to the action of fibroblasts remodelling the ECM, towhich they are attached through integrin receptors, either

by tractional forces as they migrate through the matrix orby contraction after having differentiated into myofibro-blasts [106, 124]. The process can be mimicked by cellsin collagen gels, where it has been shown that contractioninvolves the synergistic effects of both α2β1 integrin andMMPs [125], the former mediating cell-matrix interac-tions and the latter facilitating cell migration through thematrix.

Conclusion

In this review, we have described the main families ofproteases involved in wound healing and have given anoverview of their main biological functions, all of whichfall into one of the following categories: ECM synthesisand degradation; release of cryptic bioactive fragments;regulation of growth factors and cytokines; control ofcell adhesion, migration, apoptosis and signalling. Whilethe actors involved in the coordinated progression of thewound healing process are known (proteases, growth fac-tors, other cytokines, matrix proteins and their fragments),much work remains to be done to define all the possibleinteractions involved. Many missing links remain betweenobservations at the molecular and tissue levels, therebyproviding major obstacles in the way of a more integrativepicture of wound healing and other remodelling processes.Fortunately however, new experimental approaches arebeing developed that will certainly help to define the rele-vance of specific proteolytic events in particular tissuecontexts.First, new imaging techniques are opening up new possi-bilities for the direct visualisation of proteolytic activityand remodelling both in vitro and in vivo. For example,visualisation of proteolytic cell migration through collagengels has been shown to lead to re-organisation and align-ment of collagen fibres [109]. Similarly, both large scaleand local movements of ECM filaments, the latter drivenby cell migration, have been observed by time-lapse imag-ing in developing embryos [126]. In addition, the use oftimer reporters has allowed specific molecules to betracked during morphogenetic changes, as for exampleduring elastin assembly [127]. Finally, both fluorescentand radioactively labelled activity-based probes havebeen developed to directly visualise sites of proteolyticactivity in situ, as opposed to antibody-based methodswhich usually do not distinguish between precursor andactive forms. Such probes are easily designed for serineand cysteine proteases, for which numerous suicide sub-strates are known, but probes targeting metalloproteasesare also under development with interesting results [128].Also, the discovery of new protease substrates in cell-based assays has gained in efficiency with the develop-ment of quantitative proteomics and its application to pro-teases [1, 66]. Using iTRAQ (isobaric Tags for Relativeand Absolute Quantification) and SILAC (Stable IsotopeLabelling by Amino-acids in cell Culture) technologies, itis now possible, in one single experiment, to identify alarge proportion of the proteins (and their fragments)found in a defined proteome and to quantify their relativeabundances in several biological contexts (typically in thepresence and absence of a protease).Together with knock-out mice, these new developmentsshould contribute towards establishing the complex net-


work of interactions involved in each particular remodel-ling event and permit the use of this knowledge to deviseoptimized therapeutic strategies. ■Acknowledgements. We thank Daniel Kronenberg andJean-Yves Exposito for helpful comments on the manu-script. Original work from the authors’ laboratory wasfunded by grants from the European Commission(NMP2-CT-2003-504017), the Fondation Coloplast, theRégion Rhône-Alpes, the Agence National de laRecherche, the Ligue Contre le Cancer, the CNRS andthe Université Claude Bernard Lyon 1. Conflict of inter-est: none.

References

1. Auf dem Keller U, Doucet A, Overall CM. Protease research in theera of systems biology. Biol Chem 2007; 388: 1159-62.2. Doan N, Gettins PG. Human alpha2-macroglobulin is composedof multiple domains, as predicted by homology with complement com-ponent C3. Biochem J 2007; 407: 23-30.3. Lopez-Otin C, Bond JS. Proteases: Multifunctional enzymes in lifeand disease. J Biol Chem 2008; 283: 30433-7.4. Lund LR, Romer J, Bugge TH, et al. Functional overlap between twoclasses of matrix-degrading proteases in wound healing. EMBO J1999; 18: 4645-56.5. Lund LR, Green KA, Stoop AA, et al. Plasminogen activation inde-pendent of uPA and tPA maintains wound healing in gene-deficientmice. EMBO J 2006; 25: 2686-97.6. Borgono CA, Diamandis EP. The emerging roles of human tissuekallikreins in cancer. Nat Rev Cancer 2004; 4: 876-90.7. Szabo R, Bugge TH. Type II transmembrane serine proteases indevelopment and disease. Int J Biochem Cell Biol 2008; 40:1297-316.8. Conway K, Ruge F, Price P, et al. Hepatocyte growth factor regu-lation: an integral part of why wounds become chronic. WoundRepair Regen 2007; 15: 683-92.9. Lungarella G, Cavarra E, Lucattelli M, et al. The dual role of neu-trophil elastase in lung destruction and repair. Int J Biochem Cell Biol2008; 40: 1287-96.10. Korkmaz B, Moreau T, Gauthier F. Neutrophil elastase, protein-ase 3 and cathepsin G: physicochemical properties, activity andphysiopathological functions. Biochimie 2008; 90: 227-42.11. Garnero P, Borel O, Byrjalsen I, et al. The collagenolytic activityof cathepsin K is unique among mammalian proteinases. J Biol Chem1998; 273: 32347-52.12. Yasuda Y, Li Z, Greenbaum D, et al. Cathepsin V, a novel andpotent elastolytic activity expressed in activated macrophages. J BiolChem 2004; 279: 36761-70.13. Cesarman-Maus G, Hajjar KA. Molecular mechanisms of fibrino-lysis. Br J Haematol 2005; 129: 307-21.14. Grainger DJ, Wakefield L, Bethell HW, et al. Release and activa-tion of platelet latent TGF-beta in blood clots during dissolution withplasmin. Nat Med 1995; 1: 932-7.15. Sato Y, Tsuboi R, Lyons R, et al. Characterization of the activa-tion of latent TGF-beta by co-cultures of endothelial cells and pericytesor smooth muscle cells: a self-regulating system. J Cell Biol 1990;111: 757-63.16. Goldfinger LE, Stack MS, Jones JC. Processing of laminin-5 andits functional consequences: role of plasmin and tissue-type plasmino-gen activator. J Cell Biol 1998; 141: 255-65.17. Bonnefoy A, Legrand C. Proteolysis of subendothelial adhesiveglycoproteins (fibronectin, thrombospondin, and von Willebrand fac-tor) by plasmin, leukocyte cathepsin G, and elastase. Thromb Res2000; 98: 323-32.18. Li WY, Chong SS, Huang EY, et al. Plasminogen activator/plas-min system: a major player in wound healing? Wound Repair Regen2003; 11: 239-47.19. Ossowski L. guirre-Ghiso JA. Urokinase receptor and integrinpartnership: coordination of signaling for cell adhesion, migrationand growth. Curr Opin Cell Biol 2000; 12: 613-20.

20. Green KA, Lund LR. ECM degrading proteases and tissue remo-delling in the mammary gland. Bioessays 2005; 27: 894-903.21. Sterchi EE, Stocker W, Bond JS. Meprins, membrane-bound andsecreted astacin metalloproteinases. Mol Aspects Med 2008; 29:309-28.22. Stocker W, Grams F, Baumann U, et al. The metzincins--topological and sequential relations between the astacins, adamaly-sins, serralysins, and matrixins (collagenases) define a superfamily ofzinc-peptidases. Protein Sci 1995; 4: 823-40.23. Saarialho-Kere UK, Crouch EC, Parks WC. Matrix metalloprotei-nase matrilysin is constitutively expressed in adult human exocrineepithelium. J Invest Dermatol 1995; 105: 190-6.24. Ii M, Yamamoto H, Adachi Y, et al. Role of matrixmetalloproteinase-7 (matrilysin) in human cancer invasion, apoptosis,growth and angiogenesis. Exp Biol Med 2006; 231: 20-7.25. Page-McCaw A, Ewald AJ, Werb Z. Matrix metalloproteinasesand the regulation of tissue remodelling. Nat Rev Mol Cell Biol2007; 8: 221-33.26. Lee MH, Murphy G. Matrix metalloproteinases at a glance. J CellSci 2004; 117: 4015-6.27. Mott JD, Werb Z. Regulation of matrix biology by matrix metallo-proteinases. Curr Opin Cell Biol 2004; 16: 558-64.28. Newby AC. Matrix metalloproteinases regulate migration, prolif-eration, and death of vascular smooth muscle cells by degradingmatrix and non-matrix substrates. Cardiovasc Res 2006; 69:614-24.29. Nagase H, Visse R, Murphy G. Structure and function of matrixmetalloproteinases and TIMPs. Cardiovasc Res 2006; 69: 562-73.30. Porter S, Clark IM, Kevorkian L, et al. The ADAMTS metallopro-teinases. Biochem J 2005; 386: 15-27.31. Lee NV, Sato M, Annis DS, et al. ADAMTS1 mediates the releaseof antiangiogenic polypeptides from TSP1 and 2. EMBO J 2006; 25:5270-83.32. Amour A, Slocombe PM, Webster A, et al. TNF-alpha convertingenzyme (TACE) is inhibited by TIMP-3. FEBS Lett 1998; 435: 39-44.33. Amour A, Knight CG, Webster A, et al. The in vitro activity ofADAM-10 is inhibited by TIMP-1 and TIMP-3. FEBS Lett 2000; 473:275-9.34. Kashiwagi M, Tortorella M, Nagase H, et al. TIMP-3 is a potentinhibitor of aggrecanase 1 (ADAM-TS4) and aggrecanase 2 (ADAM-TS5). J Biol Chem 2001; 276: 12501-4.35. Kramerova IA, Kawaguchi N, Fessler LI, et al. Papilin in develop-ment; a pericellular protein with a homology to the ADAMTS metallo-proteinases. Development 2000; 127: 5475-85.36. Lee NV, Rodriguez-Manzaneque JC, Thai SN, et al. Fibulin-1acts as a cofactor for the matrix metalloprotease ADAMTS-1. J BiolChem 2005; 280: 34796-804.37. Tousseyn T, Jorissen E, Reiss K, et al. (Make) stick and cut loose -disintegrin metalloproteases in development and disease. BirthDefects Res C Embryo Today 2006; 78: 24-46.38. Edwards DR, Handsley MM, Pennington CJ. The ADAM metallo-proteinases. Mol Aspects Med 2008; 29: 258-89.39. Hopkins DR, Keles S, Greenspan DS. The bone morphogeneticprotein 1/Tolloid-like metalloproteinases. Matrix Biol 2007; 26:508-23.40. Ge G, Greenspan DS. BMP1 controls TGFbeta1 activation viacleavage of latent TGFbeta-binding protein. J Cell Biol 2006; 175:111-20.41. Amano S, Scott IC, Takahara K, et al. Bone morphogenetic pro-tein 1 is an extracellular processing enzyme of the laminin 5 gamma2chain. J Biol Chem 2000; 275: 22728-35.42. Rattenholl A, Pappano WN, Koch M, et al. Proteinases of thebone morphogenetic protein-1 family convert procollagen VII tomature anchoring fibril collagen. J Biol Chem 2002; 277: 26372-8.43. Gonzalez EM, Reed CC, Bix G, et al. BMP-1/Tolloid-like metal-loproteases process endorepellin, the angiostatic C-terminal fragmentof perlecan. J Biol Chem 2005; 280: 7080-7.44. Ge G, Fernandez CA, Moses MA, et al. Bone morphogeneticprotein 1 processes prolactin to a 17-kDa antiangiogenic factor.Proc Natl Acad Sci U S A 2007; 104: 10010-5.45. Lee HX, Ambrosio AL, Reversade B, et al. Embryonic dorsal-ventral signaling: secreted frizzled-related proteins as inhibitors of tol-loid proteinases. Cell 2006; 124: 147-59.46. Kobayashi K, Luo M, Zhang Y, et al. Secreted Frizzled-relatedprotein 2 is a procollagen C proteinase enhancer with a role in fibro-sis associated with myocardial infarction. Nat Cell Biol 2009; 11:46-55.47. Moali C, Font B, Ruggiero F, et al. Substrate-specific modulationof a multisubstrate proteinase. C-terminal processing of fibrillar pro-collagens is the only BMP-1-dependent activity to be enhanced byPCPE-1. J Biol Chem 2005; 280: 24188-94.


48. Scott IC, Blitz IL, Pappano WN, et al. Homologues of Twistedgastrulation are extracellular cofactors in antagonism of BMP signal-ling. Nature 2001; 410: 475-8.49. Inomata H, Haraguchi T, Sasai Y. Robust stability of the embry-onic axial pattern requires a secreted scaffold for chordin degrada-tion. Cell 2008; 134: 854-65.50. Chen BK, Leiferman KM, Pittelkow MR, et al. Localization andregulation of pregnancy-associated plasma protein a expression inhealing human skin. J Clin Endocrinol Metab 2003; 88: 4465-71.51. Glerup S, Boldt HB, Overgaard MT, et al. Proteinase inhibitionby proform of eosinophil major basic protein (pro-MBP) is a multistepprocess of intra- and intermolecular disulfide rearrangements. J BiolChem 2005; 280: 9823-32.52. Gyrup C, Oxvig C. Quantitative analysis of insulin-like growthfactor-modulated proteolysis of insulin-like growth factor bindingprotein-4 and -5 by pregnancy-associated plasma protein-A. Biochem-istry 2007; 46: 1972-80.53. Hirano M, Ma BY, Kawasaki N, et al. Mannan-binding proteinblocks the activation of metalloproteases meprin alpha and beta. J Im-munol 2005; 175: 3177-85.54. Singer AJ, Clark RA. Cutaneous wound healing. N Engl J Med1999; 341: 738-46.55. Toriseva M, Kahari VM. Proteinases in cutaneous wound heal-ing. Cell Mol Life Sci 2009; 66: 203-24.56. Eming SA, Krieg T, Davidson JM. Inflammation in wound repair:molecular and cellular mechanisms. J Invest Dermatol 2007; 127:514-25.57. Werner S, Grose R. Regulation of wound healing by growth fac-tors and cytokines. Physiol Rev 2003; 83: 835-70.58. Bergmeier W, Piffath CL, Cheng G, et al. Tumor necrosis factor-alpha-converting enzyme (ADAM17) mediates GPIbalpha sheddingfrom platelets in vitro and in vivo. Circ Res 2004; 95: 677-83.59. Tsirogianni AK, Moutsopoulos NM, Moutsopoulos HM. Woundhealing: immunological aspects. Injury 2006; 37 (Suppl 1): S5-12.60. Petri B, Bixel MG. Molecular events during leukocyte diapedesis.FEBS J 2006; 273: 4399-407.61. Gutierrez-Fernandez A, Inada M, Balbin M, et al. Increasedinflammation delays wound healing in mice deficient in collagenase-2 (MMP-8). FASEB J 2007; 21: 2580-91.62. Lin M, Jackson P, Tester AM, et al. Matrix metalloproteinase-8facilitates neutrophil migration through the corneal stromal matrix bycollagen degradation and production of the chemotactic peptide Pro-Gly-Pro. Am J Pathol 2008; 173: 144-53.63. Parks WC, Wilson CL, Lopez-Boado YS. Matrix metalloprotei-nases as modulators of inflammation and innate immunity. Nat RevImmunol 2004; 4: 617-29.64. Wolf M, Albrecht S, Marki C. Proteolytic processing of chemo-kines: implications in physiological and pathological conditions. IntJ Biochem Cell Biol 2008; 40: 1185-98.65. McQuibban GA, Gong JH, Wong JP, et al. Matrix metalloprotei-nase processing of monocyte chemoattractant proteins generates CCchemokine receptor antagonists with anti-inflammatory properties invivo. Blood 2002; 100: 1160-7.66. Dean RA, Overall CM. Proteomics discovery of metalloprotei-nase substrates in the cellular context by iTRAQ labeling reveals adiverse MMP-2 substrate degradome. Mol Cell Proteomics 2007; 6:611-23.67. Ishida Y, Gao JL, Murphy PM. Chemokine receptor CX3CR1mediates skin wound healing by promoting macrophage and fibro-blast accumulation and function. J Immunol 2008; 180: 569-79.68. Yu Q, Stamenkovic I. Cell surface-localized matrixmetalloproteinase-9 proteolytically activates TGF-beta and promotestumor invasion and angiogenesis. Genes Dev 2000; 14: 163-76.69. Taipale J, Keski-Oja J. Growth factors in the extracellular matrix.FASEB J 1997; 11: 51-9.70. Melrose J, Hayes AJ, Whitelock JM, et al. Perlecan, the “jack ofall trades” proteoglycan of cartilaginous weight-bearing connectivetissues. Bioessays 2008; 30: 457-69.71. Bergers G, Brekken R, McMahon G, et al. Matrixmetalloproteinase-9 triggers the angiogenic switch during carcino-genesis. Nat Cell Biol 2000; 2: 737-44.72. Roy R, Zhang B, Moses MA. Making the cut: protease-mediatedregulation of angiogenesis. Exp Cell Res 2006; 312: 608-22.73. Lee S, Jilani SM, Nikolova GV, et al. Processing of VEGF-A bymatrix metalloproteinases regulates bioavailability and vascular pat-terning in tumors. J Cell Biol 2005; 169: 681-91.

74. Dean RA, Butler GS, Hamma-Kourbali Y, et al. Identification ofcandidate angiogenic inhibitors processed by matrix metalloprotei-nase 2 (MMP-2) in cell-based proteomic screens: disruption of vascu-lar endothelial growth factor (VEGF)/heparin affin regulatory peptide(pleiotrophin) and VEGF/Connective tissue growth factor angiogenicinhibitory complexes by MMP-2 proteolysis. Mol Cell Biol 2007; 27:8454-65.75. Bunn RC, Fowlkes JL. Insulin-like growth factor binding proteinproteolysis. Trends Endocrinol Metab 2003; 14: 176-81.76. Rifkin DB. Latent transforming growth factor-beta (TGF-beta) bind-ing proteins: orchestrators of TGF-beta availability. J Biol Chem2005; 280: 7409-12.77. Isogai Z, Ono RN, Ushiro S, et al. Latent transforming growthfactor beta-binding protein 1 interacts with fibrillin and is amicrofibril-associated protein. J Biol Chem 2003; 278: 2750-7.78. Neptune ER, Frischmeyer PA, Arking DE, et al. Dysregulation ofTGF-beta activation contributes to pathogenesis in Marfan syndrome.Nat Genet 2003; 33: 407-11.79. Wipff PJ, Hinz B. Integrins and the activation of latent transform-ing growth factor beta1 - an intimate relationship. Eur J Cell Biol2008; 87: 601-15.80. Mu D, Cambier S, Fjellbirkeland L, et al. The integrin alpha(v)beta8 mediates epithelial homeostasis through MT1-MMP-dependentactivation of TGF-beta1. J Cell Biol 2002; 157: 493-507.81. Schultz-Cherry S, Murphy-Ullrich JE. Thrombospondin causes acti-vation of latent transforming growth factor-beta secreted by endothe-lial cells by a novel mechanism. J Cell Biol 1993; 122: 923-32.82. Ribeiro SM, Poczatek M, Schultz-Cherry S, et al. The activationsequence of thrombospondin-1 interacts with the latency-associatedpeptide to regulate activation of latent transforming growth factor-beta. J Biol Chem 1999; 274: 13586-93.83. Wipff PJ, Rifkin DB, Meister JJ, et al. Myofibroblast contractionactivates latent TGF-beta1 from the extracellular matrix. J Cell Biol2007; 179: 1311-23.84. Higashiyama S, Nanba D. ADAM-mediated ectodomain shed-ding of HB-EGF in receptor cross-talk. Biochim Biophys Acta 2005;1751: 110-7.85. Nanba D, Mammoto A, Hashimoto K, et al. Proteolytic release ofthe carboxy-terminal fragment of proHB-EGF causes nuclear export ofPLZF. J Cell Biol 2003; 163: 489-502.86. Bray SJ. Notch signalling: a simple pathway becomes complex.Nat Rev Mol Cell Biol 2006; 7: 678-89.87. Chigurupati S, Arumugam TV, Son TG, et al. Involvement ofNotch signaling in wound healing. PLOS One 2009; 2: e1167.88. Murphy G, Murthy A, Khokha R. Clipping, shedding and RIPpingkeep immunity on cue. Trends Immunol 2008; 29: 75-82.89. Chambers RC, Laurent GJ. Coagulation cascade proteases andtissue fibrosis. Biochem Soc Trans 2002; 30: 194-200.90. Jenkins G. The role of proteases in transforming growth factor-beta activation. Int J Biochem Cell Biol 2008; 40: 1068-78.91. Romer J, Bugge TH, Pyke C, et al. Impaired wound healing inmice with a disrupted plasminogen gene. Nat Med 1996; 2:287-92.92. Hintermann E, Quaranta V. Epithelial cell motility on laminin-5:regulation by matrix assembly, proteolysis, integrins and erbB recep-tors. Matrix Biol 2004; 23: 75-85.93. Remy L, Trespeuch C, Bachy S, et al. Matrilysin 1 influencescolon carcinoma cell migration by cleavage of the laminin-5 beta3chain. Cancer Res 2006; 66: 11228-37.94. Dwivedi A, Slater SC, George SJ. MMP-9 and -12 cause N-cadherin shedding and thereby beta-catenin signalling and vascularsmooth muscle cell proliferation. Cardiovasc Res 2009; 81: 178-86.95. Noe V, Fingleton B, Jacobs K, et al. Release of an invasion pro-moter E-cadherin fragment by matrilysin and stromelysin-1. J Cell Sci2001; 114: 111-8.96. Maretzky T, Reiss K, Ludwig A, et al. ADAM10 mediates E-cadherin shedding and regulates epithelial cell-cell adhesion, migra-tion, and beta-catenin translocation. Proc Natl Acad Sci U S A 2005;102: 9182-7.97. Mayerle J, Schnekenburger J, Kruger B, et al. Extracellular cleav-age of E-cadherin by leukocyte elastase during acute experimentalpancreatitis in rats. Gastroenterology 2005; 129: 1251-67.98. Huguenin M, Muller EJ, Trachsel-Rosmann S, et al. The metallo-protease meprinbeta processes E-cadherin and weakens intercellularadhesion. PLoS ONE 2008; 3: e2153.99. George SJ, Dwivedi A. MMPs, cadherins, and cell proliferation.Trends Cardiovasc Med 2004; 14: 100-5.100. Tran KT, Lamb P, Deng JS. Matrikines and matricryptins: Impli-cations for cutaneous cancers and skin repair. J Dermatol Sci 2005;40: 11-20.


101. Bix G, Iozzo RV. Matrix revolutions: “tails” of basement-membrane components with angiostatic functions. Trends Cell Biol2005; 15: 52-60.102. Stringa E, Knauper V, Murphy G, et al. Collagen degradationand platelet-derived growth factor stimulate the migration of vascularsmooth muscle cells. J Cell Sci 2000; 113 (Pt 11): 2055-64.103. Xu J, Rodriguez D, Petitclerc E, et al. Proteolytic exposure of acryptic site within collagen type IV is required for angiogenesis andtumor growth in vivo. J Cell Biol 2001; 154: 1069-79.104. O’Reilly MS, Wiederschain D, Stetler-Stevenson WG, et al.Regulation of angiostatin production by matrix metalloproteinase-2in a model of concomitant resistance. J Biol Chem 1999; 274:29568-71.105. Mannello F, Luchetti F, Falcieri E, et al. Multiple roles of matrixmetalloproteinases during apoptosis. Apoptosis 2005; 10: 19-24.106. Rhee S, Grinnell F. Fibroblast mechanics in 3D collagen matri-ces. Adv Drug Deliv Rev 2007; 59: 1299-305.107. Hinz B. Formation and function of the myofibroblast during tis-sue repair. J Invest Dermatol 2007; 127: 526-37.108. Zaman MH, Trapani LM, Sieminski AL, et al. Migration of tumorcells in 3D matrices is governed by matrix stiffness along with cell-matrix adhesion and proteolysis. Proc Natl Acad Sci U S A 2006;103: 10889-94.109. Friedl P, Wolf K. Proteolytic interstitial cell migration: a five-stepprocess. Cancer Metastasis Rev, 2009.110. Alexander NR, Branch KM, Parekh A, et al. Extracellular matrixrigidity promotes invadopodia activity. Curr Biol 2008; 18: 1295-9.111. Hulmes DJS. Collagen diversity, synthesis and assembly. In:Fratzl P, ed. Collagen - Structure and Mechanics. New York:Springer, 2008: 15-47.112. Kadler KE, Hill A, Canty-Laird EG. Collagen fibrillogenesis:fibronectin, integrins, and minor collagens as organizers and nuclea-tors. Curr Opin Cell Biol 2008; 20: 495-501.113. Gopalakrishnan B, Wang WM, Greenspan DS. Biosyntheticprocessing of the Pro-alpha1(V)Pro-alpha2(V)Pro-alpha3(V) procolla-gen heterotrimer. J Biol Chem 2004; 279: 30904-12.114. Scott IC, Imamura Y, Pappano WN, et al. Bone morphogeneticprotein-1 processes probiglycan. J Biol Chem 2000; 275: 30504-11.115. Ge G, Seo NS, Liang X, et al. Bone morphogenetic protein-1/tolloid-related metalloproteinases process osteoglycin and enhanceits ability to regulate collagen fibrillogenesis. J Biol Chem 2004; 279:41626-33.

116. Lucero HA, Kagan HM. Lysyl oxidase: an oxidative enzymeand effector of cell function. Cell Mol Life Sci 2006; 63: 2304-16.117. Telci D, Griffin M. Tissue transglutaminase (TG2) - a woundresponse enzyme. Front Biosci 2006; 11: 867-82.118. Uzel MI, Scott IC, Babakhaniou-Chase H, et al. Multiple bonemorphogenetic protein 1-related mammalian metalloproteinases pro-cess pro-lysyl oxidase at the correct physiological site and controllysyl oxidase activation in mouse embryo fibroblast cultures. J BiolChem 2001; 276: 22537-43.119. Atsawasuwan P, Mochida Y, Katafuchi M, et al. Lysyl oxidasebinds transforming growth factor-beta and regulates its signaling viaamine oxidase activity. J Biol Chem 2008; 283: 34229-40.120. Metcalfe AD, Ferguson MW. Tissue engineering of replace-ment skin: the crossroads of biomaterials, wound healing, embryonicdevelopment, stem cells and regeneration. J R Soc Interface 2007; 4:413-37.121. Caplan AI. Embryonic development and the principles of tissueengineering. Novartis Found Symp 2003; 249: 17-25.122. Shah M, Foreman DM, Ferguson MWJ. Neutralising antibodyto TGF-b1,2 reduces cutaneous scarring in adult rodents. J Cell Sci1994; 107: 1137-57.123. Connon CJ, Meek KM. Organization of corneal collagen fibrilsduring the healing of trephined wounds in rabbits. Wound RepairRegen 2003; 11: 71-8.124. Tomasek JJ, Gabbiani G, Hinz B, et al. Myofibroblasts andmechano-regulation of connective tissue remodelling. Nat Rev MolCell Biol 2002; 3: 349-63.125. Phillips JA, Bonassar LJ. Matrix metalloproteinase activity syner-gizes with alpha2beta1 integrins to enhance collagen remodeling.Exp Cell Res 2005; 310: 79-87.126. Czirok A, Zamir EA, Filla MB, et al. Extracellular matrix macro-assembly dynamics in early vertebrate embryos. Curr Top Dev Biol2006; 73: 237-58.127. Kozel BA, Rongish BJ, Czirok A, et al. Elastic fiber formation: adynamic view of extracellular matrix assembly using timer reporters.J Cell Physiol 2006; 207: 87-96.128. Blum G. Use of fluorescent imaging to investigate pathologicalprotease activity. Curr Opin Drug Discov Devel 2008; 11: 708-16.


Documents

Catherine MOALI Extracellular and cell surface proteases ......Wound healing is a complex example of tissue remodelling that includes several steps occurring either concomitantly or