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

ARTICLE

Auteur(s) : Catherine Moali, David JS Hulmes

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

Tissue remodelling, the acquisition of new tissue architecture resulting from the combined action of cells and their extracellular matrix (ECM), is a feature of normal development and growth, as well as pathological situations and tissue repair. In most tissues, the amount of ECM equals or exceeds the cellular component and most of the ECM is synthesized by mesenchymal cells (fibroblasts, myofibroblasts and smooth muscle cells). Epithelial and endothelial cells also contribute by synthesizing specialised ECM structures such as basement membranes, which have both functional and protective roles. Exactly how cells orchestrate the organisation of the ECM remains largely unclear but extracellular and cell surface proteases have clearly been identified as key players due to their ability to synchronize matrix synthesis and degradation with the activation and deactivation of growth factors and other cytokines.

While normal tissue homeostasis depends on a controlled balance between synthesis and degradation, loss of homeostasis in diseases of the skin, cornea or internal organs can lead either to excessive production (e.g. fibrosis, hypertrophic scarring) or excessive destruction (e.g. arthritis, emphysema, chronic ulcers) of ECM. Until recently, remodelling was regarded almost exclusively from the destructive point of view with little attention being paid to synthesis. This is explained by the intensive research for new anti-tumoral agents and by the great expectations raised by inhibitors of matrix metalloproteinases (MMPs). Even now, the remodelling literature only rarely mentions proteases involved in ECM synthesis and one major aspect of this review is to recognize the importance of proteases such as tolloids and ADAMTSs in the maturation of matrix components.

In recent years, using knock-out mice and novel tools such as activity-based probes and high-throughput proteomics, researchers have achieved real progress towards understanding the “protease web” [1] which links proteases to other proteases, to their substrates, their inhibitors and their various binding partners. However, it remains a challenge to link substrates with observations at the cellular or tissue level, and in vivo extrapolation of processing events observed in vitro is made difficult by the intrinsic complexity of the extracellular matrix. Substrates can be masked by interaction partners, they can adopt different conformations, or be physically entrapped or otherwise inaccessible to proteases, due to spatial or temporal separation.

In contrast to most other post-translational modifications, proteolytic events are irreversible and therefore fine tuning of protease activity is required to avoid tissue destruction. For all types of extracellular proteases, inhibitors have 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 for enhancement of proteolytic activity have also been described and, interestingly, these usually provide a means to direct protease activity to particular substrates. Other regulatory mechanisms following secretion include propeptide removal, feedback loops involving processed substrates and control of extracellular localisation.

In this review, we give examples of these various aspects and present an overview of the properties and functions of the extracellular and cell surface proteases (mainly serine proteases and metalloproteases) involved in wound healing. This well known case of tissue remodelling illustrates all the different functions with which extracellular and cell surface proteases are known to be involved: matrix protein synthesis and degradation; release of cryptic bioactive fragments; regulation of growth factors and other cytokines; and the control of cell adhesion, migration, apoptosis and signalling.

Extracellular and cell surface proteases: old and new players

Each protease family could be the subject of an article in its own right; here we present the main features of these enzymes and refer the reader to the numerous reviews now available for more detailed information.

Serine and cysteine proteases

More broadly relevant in the wound healing context are proteases from the plasmin system, which are not only responsible for fibrin clot lysis [13] but can also participate in TGF-β activation [14, 15], cleave components of the extracellular matrix (laminin 332 [16], thrombospondin-1, fibronectin [17]) or activate other proteases (especially MMPs) which will in turn degrade ECM. In addition, as mentioned above, the crucial role of plasmin in tissue repair has been demonstrated in mice deficient in plasminogen [4, 18].

Figure 1 illustrates the complex regulation of plasminogen conversion into the active plasmin form [13]. Plasmin activation can be seen as a paradigm for protease regulation with both enhancing and inhibiting mechanisms both contributing to the fine tuning of protease activity. The conversion of plasminogen to plasmin is mainly carried out by two other serine proteases: tPA (tissue-type plasminogen activator) and uPA (urokinase-type plasminogen activator) with the latter being the most important for tissue remodelling and the former for fibrin clot dissolution.

Several modulators of plasmin, tPA and uPA have been described that, in some cases, can act in a substrate-specific manner. Such substrate-specific regulation is frequently encountered with proteases, but rarely with other types of enzymes. For example, fibrin, the major plasmin substrate, binds both tPA and plasminogen, thereby increasing tPA catalytic efficiency and enhancing fibrinolysis [13]. In addition, uPA can be recruited and activated more efficiently at the cell surface via a specific receptor uPAR (uPA receptor). uPAR can then associate with a variety of cell receptors, including integrins, and modulate cell adhesion [19, 20]. Through a positive feedback loop, plasmin cleaves tPA and uPA and the resulting two-chain plasminogen activators are more active than the single chain proteases. Finally, two inhibitors of the plasminogen activators (PAI-1 and PAI-2) and inhibitors of plasmin (such as α2-plasmin inhibitor) further regulate plasmin activation. This results in a complex pathway of plasmin maturation, as described in figure 1, accounting for the destructive potential associated with this protease.

Metalloproteases

Matrix metalloproteases (MMPs)

The initial classification of the 23 human MMPs according to their substrate specificity and functional characteristics (collagenases, gelatinases, broad-specificity stromelysins, membrane-bound MMPs…) has now given way to a simple numbering scheme. While all MMPs have in common a signal peptide, a propeptide and a catalytic domain, they differ by the possible insertion of fibronectin-type II repeats in the catalytic domain (MMPs-2 and -9), the presence of a haemopexin domain involved in substrate recognition (in all MMPs except MMPs-7 and -26), an additional transmembrane domain followed by a cytoplasmic tail (MMPs-14, -15, -16 and 24) or a GPI anchor (MMPs-17 and -25) [26]. The propeptide is cleaved off either in the Golgi apparatus by proteases from the furin-like proprotein convertase family or pericellularly by plasmin, other MMPs or by autoproteolysis.

As a family, MMPs can cleave all extracellular matrix components (collagens, elastin, proteoglycans and other structural glycoproteins). They are also involved in growth factor activation (e.g. TGF-β and VEGF: see below) and can modify cell-matrix and cell-cell interactions in several ways (including matrix degradation, shedding of syndecans and cadherins, and binding of integrins and CD44) [27, 28]. After propeptide removal, MMP activity is regulated by potent inhibitors of the TIMP family (Tissue Inhibitors of MMPs 1-4) which have also been implicated in a range of biological processes, grossly overlapping MMP functions but with opposite effects: tumour metastasis, angiogenesis, wound healing, cell differentiation and survival, inflammatory responses, etc. TIMPs bind to the catalytic zinc ion of MMPs, thus blocking access to the catalytic pocket. All MMPs can be inhibited by all TIMPs but to variable extents [29]. The main structural and regulatory features of this family are also summarized in figure 1.

ADAMs and ADAMTSs

With a few exceptions, ADAM proteins remain attached to the plasma membrane and the catalytic activity of the ADAMs that are predicted to be active (8-10, 12, 15, 17, 19-21, 28, 30, 33) is devoted to the shedding of the ectodomain of several transmembrane proteins. They release diffusible factors, such as heparin-binding EGF (HB-EGF) or TNF-α, from their membrane-anchored precursors. They also cleave adhesion molecules (e.g. cadherins and CD44) to disrupt adhesive contacts while shedding of signalling receptors (e.g. the Notch receptor) leads to further intra-membrane proteolysis and release of an intracellular domain which becomes a signal transducer (see below). In the latter case, binding of a substrate ligand, such as Delta in the case of the Notch receptor, is often required to trigger proteolysis by ADAMs and permits very strict regulation of the downstream signalling pathway. The ADAMs that are proteolytically inactive can nevertheless play important roles, especially in cell adhesion, via their disintegrin and cysteine-rich domains. The disintegrin domains of several ADAMs, for example, have been shown to bind to receptors of the integrin family and thereby block their function. Furthermore, the cysteine-rich domain can interact with ECM proteins such as fibronectin and heparan-sulphate proteoglycans of the syndecan family.

In contrast to the ADAMs, ADAMTSs are all secreted proteases and are found either in soluble form or bound to the ECM. Their key feature is a variable number of type 1 thrombospondin repeats (figure 1). They are all supposedly active but six of the ADAMTS are orphans with no known substrates. Other ADAMTSs tend to be much more specific than the metalloproteases described so far and their main remodelling functions can be classified into 3 overlapping groups [30]: (i) the anti-angiogenic ADAMTSs (-1 and -8), this function being directly mediated 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 during joint degeneration and (iii) the procollagen N-proteinases (-2, -3 and -14) which are involved in processing of fibrillar collagen precursors.

For both ADAMs and ADAMTSs, propeptide removal is carried out by proteases of the proprotein convertase family in the Golgi apparatus. The only protein inhibitors that have been described so far for ADAMs are the TIMPs but in contrast to the MMPs, not all members of the family are inhibited by the four TIMPs and the physiological consequences of this inhibition are not very well documented. TIMP-3, which is unique among the TIMPs for its ability to interact strongly with ECM, is also a potent inhibitor 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 of ADAMTS-2 but this particular enzyme has been shown to be inhibited by papilin [35], a protein found in the basement membrane which contains the C-terminal, non catalytic domains of ADAMTSs. Finally, ADAMTS-1-dependent cleavage of aggrecan seems to be enhanced by fibulin-1 through simultaneous binding of both protease and substrate [36].

For a more comprehensive overview of the members and roles 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

From a structural point of view, tolloid proteinases comprise a catalytic domain of the astacin type and a variable number of CUB and EGF domains (figure 1). Their propeptides are also removed by furin-like proteases, as for some MMPs and almost all ADAM(TS)s. Tolloids are soluble proteins which are regulated substrate-specific enhancers. In addition, in Xenopus, it has been shown that the Sizzled protein binds strongly to tolloids and inhibits cleavage of chordin [45], suggesting that homologous secreted frizzled-related proteins might play similar roles in mammals (though not supported by recent data, [46]). Tolloids are also regulated by enhancer proteins which are substrate-specific in that they specifically activate the cleavage of only one type of substrate, for example procollagen C-proteinase enhancers (PCPEs) in the case of the fibrillar procollagen substrates [47] or twisted gastrulation and olfactomedin in the case of chordin [48, 49].

Other metalloproteases

Meprins, like tolloid enzymes, are proteases of the astacin family [21], which, unlike tolloid proteases, are synthesised as transmembrane proteins. They are expressed as two independent subunits (α and β) which can assemble into hetero- or homo-dimers, or higher oligomers in the case of meprin α, which can be released from the plasma membrane by proteolytic shedding. Expression of meprins has been reported mainly in the kidney, intestine and skin but could increase during inflammatory conditions, cancer progression and wound healing. In recent years, meprins have been shown to cleave several ECM proteins (collagen IV, laminin 111, nidogen-1, fibronectin), enzymes such as lysyl oxidases and adhesion proteins such as cadherins but the discovery of physiologically relevant substrates is still a matter of intense research. The only endogenous inhibitor described so far for meprins is mannan-binding lectin [53] but the primary regulatory mechanism of meprins after propeptide removal seems to be targeting to specific locations (e.g. apical membrane of polarized cells).

Extracellular and cell surface proteases in wound healing

During the first phase, haemostasis and inflammation, a fibrin-rich provisional matrix is formed as a result of thrombin and platelet activation. Fragments of coagulation factors, activated complement components, growth factors and chemokines released by damaged cells and activated platelets then recruit inflammatory cells to the site of injury [54, 55]. Platelets, for example, are major sources of TGF-β and PDGF. First to arrive are the neutrophils, which cleanse the wound area by phagocytosis and the release of antimicrobial substances (reactive oxygen species, cationic peptides, eicosanoids) and proteases (neutrophil elastase, cathepsin G, proteinase 3, uPA, MMPs-8 and -9) [56]. Migration of neutrophils from the blood capillaries into the wound site involves interactions with endothelial cell adhesion molecules such as selectins and ICAMs. Subseqently, monocytes also infiltrate the wound site, in response to growth factors (TGF-β, PDGF) and chemokines (e.g. monocyte chemoattractant proteins or MCPs) where they become activated and differentiate into mature tissue macrophages. While both neutrophils and macrophages produce pro-inflammatory mediators (TNF-α, interleukins-1 and -6), macrophages also synthesize several growth factors (TGF-β, TGF-α, bFGF, PDGF, VEGF) to stimulate the repair response.

In the second phase, re-epithelialisation and granulation tissue formation, which commences soon after injury, keratinocytes use the provisional fibrin-rich matrix to migrate and cover the wound area. This requires the coordinated action of several growth factors and cytokines, together with multiple proteases and adhesion proteins involved in cell-cell and cell-matrix interactions. Important growth factors involved in re-epithelialisation include members of the EGF family (EGF, TGF-α, HB-EGF) and FGFs, which promote keratinocyte proliferation, and TGF-β, which negatively regulates this process [57]. Next, about 4 days after injury, fibroblasts are also attracted to the wound area to initiate the formation of granulation tissue, which includes a new extracellular matrix rich in collagen III, fibronectin and proteoglycans. Cross-talk between keratinocytes and fibroblasts plays an important role, notably in the production of a new basement membrane, as well as in the control of cell proliferation. Fibroblast proliferation is stimulated by different growth factors (including PDGF, TGF-β, CTGF, IGFs) some of which also promote the important process of wound contraction by stimulating fibroblast differentiation into myofibroblasts. In addition to the acquisition of a contractile phenotype, these myofibroblasts also demonstrate an enhanced ability to synthesize extracellular matrix. Finally, formation of granulation tissue also requires new blood vessel formation, angiogenesis, under the control of different growth factors (notably VEGF) as well as anti-angiogenic factors released by proteolysis (see below).

In the final phase of tissue repair, tissue remodelling, the granulation tissue is progressively replaced by a more mature scar tissue, where collagen I is the major component. Remodelling is a lengthy process, which requires both synthesis and degradation of ECM in order to recreate as closely as possible, but never completely, the properties of the original tissue. At the end of wound repair, myofibroblasts usually undergo apoptosis and are replaced by normal fibroblasts. However, in certain pathologies, excessive inflammation can lead to persistence of myofibroblasts and excessive ECM synthesis, leading to hypertrophic scarring, or complete failure to heal, as in chronic wounds (ulcers).

Proteases play numerous roles during the process of wound healing (table 1). They are at work during: haemostasis; regulation of the activities of growth factors, chemokines and other cytokines; cell migration, signalling and survival; angiogenesis; and synthesis and remodelling of the extracellular matrix.

Haemostasis and inflammation

Circulating leukocytes then infiltrate into the wound area first by attachment to the endothelial cells of the blood vessel wall followed by extravasation into the surrounding tissue. This involves tethering to specific cell adhesion molecules expressed on the endothelial cells, notably E-selectin, ICAM-1 and VCAM-1, whose expression is induced by inflammatory cytokines [59]. Subsequent extravasation through the blood vessel wall requires additional 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 well as by shedding of L-selectin from neutrophils. Finally, neutrophil infiltration into the wound site is facilitated by secretion of the neutrophil collagenase MMP-8, probably through collagen degradation but also through proteolytic modification of chemotactic molecules [61, 62].

Migration of inflammatory cells into the wound site is actually controlled by several chemokines, many of which are regulated by the action of proteases [63, 64]. Protease activity can lead to increased or decreased activity, either directly, by proteolytic cleavage of the chemokine, or indirectly, by cleavage of interaction partners or receptors. Particularly well studied are the monocyte chemoattractant proteins (MCPs) which serve to attract monocytes to the site of tissue injury. A number of MMPs have been found to cleave short N-terminal fragments from different MCPs leading to a loss of chemokine activity and the generation of receptor antagonists [65]. Since these MMPs are produced by stromal cells as a result of macrophage-derived inflammatory cytokines, such activity could serve to dampen the inflammatory response. Alternatively, other chemokines can be directly activated by proteolytic activity, either by the processing of inactive precursors (such as IL-8/CXCL8 activation by serine proteases, cathepsin L or MMP-9) or by shedding of membrane-bound precursors (such as CX3CL1/fractalkine by MMP-2 [66], ADAMs-10 and -17 [67]). Furthermore, since IL-8 binds to the heparan sulphate chains of cell surface syndecan-1, MMP-7 shedding of the IL-8/syndecan-1 complex can help establish local chemokine gradients at the wound site [63].

Finally, ADAM-17/TACE also plays a key role in the regulation of proinflammatory cytokines, both by the shedding of TNF-α from the cell surface, as well as by the shedding of both TNF-α and IL-6 receptors [38]. Similarly, MMPs-7 and -12 may also release TNF-α from macrophages, while MMPs-2,-3 and -9 are involved in the activation and inactivation of IL-1β [63].
Table 1 Biological functions of extracellular and cell surface proteases in tissue remodelling: substrates and proteases involved. Compiled with the help of the CutDB proteolytic event database (http://cutdb.burnham.org/), from where original references can be obtained

Re-epithelialisation, angiogenesis and granulation tissue formation

Many of the examples described below show the importance of proteolytic processing at the cell surface, not only for transmembrane proteases (ADAMs, meprins, membrane type MMPs, matriptases) but also for other proteases that are anchored to the plasma membrane, either directly, through a GPI anchor (MMPs-17,-25), or indirectly, in association with cell surface receptors. Examples of the latter are binding of MMP-9 to the hyaluronan receptor CD44 [68], MMP-2 to integrin α_vβ₃, MMP-1 to integrin α₂β₁ and MMP-7 to heparan sulphate proteoglycans [63]. Such interactions serve to localise the site of proteolysis, or protease activation, as well as to bring proteases and their substrates in closer proximity.

Release of growth factors sequestered in the extracellular matrix

The activation of TGF-β has been extensively studied, and several mechanisms, both proteolytic and non-proteolytic, have been elucidated. TGF-β, (of which there are three variants β1, β2 and β3), is secreted as a latent dimeric complex consisting of the functional growth factor at the C-terminus and the latent prodomain (called LAP or latency-associated peptide) at the N-terminus. TGF- β is cleaved from its propeptide during secretion by furin-like proteases and remains bound to LAP in a non-covalent manner. LAP can also form disulphide bridges with latent TGF- β binding proteins LTBPs-1, -3 or -4 [76] to produce LLC (large latent complex). Because LTBPs have a strong affinity for several matrix components, especially fibrillin-1 [77, 78], LLC is deposited within the extracellular matrix, to which it is covalently bound by transglutaminase cross-linking [79]. Release of the TGF-β/LAP complex from the ECM requires proteolysis of the LTBP and this processing was recently attributed to tolloids [40] (Fig.2). Subsequently, active TGF-β can be released from the complex by another round of proteolysis, this time under the control of plasmin [14] or MMPs-2, -9 and -14, themselves possibly in association with integrins, to which LAP also binds [68, 79, 80]. Conformational changes leading to TGF-β activation in the absence of proteolytic activity have also been demonstrated, the first being mediated by thrombospondin-1 (TSP-1) which can bind to LAP associated with TGF-β [81]. This interaction appears to induce a conformational change in the prodomain which, while remaining bound to mature TGF-β, leads to a loss of latency and hence TGF-β activation [82]. More recently, it has been shown that mechanical forces exerted by myofibroblasts attached to an extracellular matrix substrate can directly lead to TGF-β activation, through an LTBP-dependent but non-proteolytic pathway [83]. With LAP bound to different types of integrins via its RGD motif, cell shape change or movement can distort the structure of the latent complex, thereby releasing active TGF-β [79]. Interestingly, this mechanism only works when cells remain bound to a relatively rigid substrate, in order to provide sufficient mechanical traction, thereby demonstrating the importance of the physical properties of the extracellular matrix in growth factor activation.

Release of membrane-bound growth factors and control of cell signalling

In addition to growth factor activation, shedding also plays a role in cell signalling. One of the best studied examples is the Notch signalling pathway which is involved in a large panel of developmental events and in several cancers [86] and was recently shown to be directly involved in wound healing [87]. Thus, when Notch signalling is blocked, delayed healing is observed, due to changes in endothelial cell, keratinocyte and fibroblast migration. Both the Notch receptor and its ligands (members of the Delta and Jagged families) are transmembrane proteins with large extracellular domains and as a consequence, interaction between receptor and ligands is restricted to neighbouring cells. When a ligand is bound to the Notch receptor, two sequential proteolytic events are induced: the first cleavage is performed by ADAM-10 or -17 resulting in the release of the Notch extracellular domain, while the second cleavage occurs inside the membrane, by regulated intracellular proteolysis or RIPping, and is due to the γ-secretase complex [88]. The second cleavage promotes the translocation of the Notch intracellular domain into the nucleus with subsequent induction of transcriptional events.

Another example of the importance of proteases in cell signalling has been described for the family of protease activated receptors (PARs). These are transmembrane G-protein coupled receptors in which cleavage of part of the extracellular domain (usually by serine proteases) exposes a tethered ligand (part of the polypeptide chain) which then interacts with the rest of the molecule, thereby triggering intracellular signalling in an irreversible manner. Activation of PARs by thrombin can have a direct effect on the expression of profibrotic mediators such as PDGF and CTGF [89], as well as TGF-β [90] and has been shown to be involved in tissue fibrosis.

Modulation of cell-matrix and cell-cell interactions

With regard to cell-cell interactions, those mediated by cadherins (E-cadherin in the case of keratinocytes or VE-cadherin the case of endothelial cells) are also subject to protease regulated shedding. Several proteases are potentially 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 translocation 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 cryptic sites or fragments (matricryptins) that change cell-matrix interactions [100, 101]. For example, PDGF-stimulated vascular smooth muscle cells migrate more rapidly on collagen I fragments generated by collagenase activity than on intact collagen [102]. This is due to changes in adhesion and migration mediated by a switch in integrin recognition from α₂β₁ to α_vβ₃. Similarly, MMP cleavage of basement membrane collagen IV exposes a cryptic site associated with increased angiogenesis coupled with decreased binding to integrin α₁β₁ and increased binding to αvβ3 [103]. Further examples are the release of anti-angiogenic fragments from pro-angiogenic molecules or even from proteases themselves, including: (i) angiostatin from plasmin [104], (ii) anti-angiogenic fragments from basement membrane components, including endostatin from collagen XVIII, tumstatin, canstatin and arresten from collagen IV and endorepellin from perlecan [43, 101], and (iii) anti-angiogenic fragments from prolactin and growth hormone [44] (figure 2).

Finally, in addition to their effects on cell proliferation, signalling, adhesion and migration, extracellular proteases are also known to influence programmed cell death, apoptosis [105]. Proteolytic disruption of cell-matrix interactions can lead to apoptosis by the process of anoikis. Alternatively, apoptosis can be induced by shedding of Fas ligand, or inhibited by release of its “death” receptor Fas/CD95, both of which have been reported for MMP-7 [24].

Matrix remodelling

Granulation tissue formation involves the relatively rapid deposition of fibrillar collagens (mainly collagen III), fibronectin, proteoglycans and other proteins in a newly synthesized ECM. The tolloids are the group of extracellular proteases that have been most clearly linked to the process of ECM assembly [39] (figure 2). Fibrillar collagens (types I, II, III, V and XI) are synthesized in precursor form, with each molecule having both N- and C-terminal propeptide extensions [111, 112]. The propeptides help maintain solubility during synthesis of the procollagen molecule and are removed either just before or shortly after secretion into the ECM. In the case of the major fibrillagen collagens (types I, II and III), the C-propeptides are removed by tolloid proteases, while the N-propeptides are removed by members of the ADAMTS family (types -2, -3 and -14). Following cleavage of propeptides, mature collagen molecules can then spontaneously assemble into fibrils. Proteolytic processing of procollagens V and XI is more complicated, with tolloids also being involved in the cleavage of N-propeptides and furin-like enzymes contributing to C-terminal maturation [113]. Most collagen fibrils are usually heterotypic in nature, i.e. composed of different collagen types, for examples collagen I/III/V fibrils in skin or collagen II/XI fibrils in cartilage, and it is known that such heterotypic interactions play a role in the regulation of collagen fibril diameter [111]. In addition, variations in the rates and extent of processing between different collagen types, particularly in the N-terminal regions, also contribute to diameter regulation. Collagen fibril assembly is further regulated by interactions with members of the small leucine rich proteoglycans (SLRPs, including decorin, fibromodulin, biglycan, osteoglycin), among which biglycan and osteoglycin have been shown to be cleaved by tolloid proteases [114, 115]. In the case of osteoglycin, the tolloid-dependent processing potentiates inhibition of collagen fibril assembly. Once assembled, collagen fibrils are stabilised by the formation of covalent crosslinks between neighbouring molecules, in a process catalysed by cross-linking enzymes such as lysyl oxidases and transglutaminases [116, 117]. Lysyl oxidases are themselves synthesized in a precursor form, and can be activated by tolloid protease cleavage of propeptides [118]. In addition, lysyl oxidases also act on elastin fibres as well as in bFGF and TGF-β signalling [119]. With regard to the basement membrane, tolloid proteases regulate the assembly of anchoring filaments, involved in stabilizing the dermal-epidermal junction, by propeptide cleavage of the non-fibrillar procollagen VII [42], as well as incorporation of laminin 332, by processing of the γ2 chain [41] (figure 2). Bearing in mind that tolloid proteases also catalyze the first step of TGF-β release from the matrix and that TGF-β increases the expression of most matrix components, tolloid proteases appear therefore as major hubs in the control of ECM assembly.

Conversion of granulation tissue into the mature form is a process that can take several months and rarely results in the same structural and functional properties as those of the original tissue [54]. Remodelling requires the coordinated action of numerous proteases involved in ECM degradation and synthesis (essentially MMPs and ADAMTSs for the degradative part; table 1) as well as in the control of cell proliferation, differentiation, adhesion, migration and death, through the various mechanisms described above. In adults, the end result of the remodelling phase is often a scar [120] which reflects differences in the processes of repair and regeneration, where the former has evolved as a means to escape danger by rapid recovery of tissue integrity [121]. Scars are characterized by dense bundles of mostly type I collagen fibrils which appear to form during the process of wound contraction and become oriented along lines of mechanical stress [122, 123]. Wound contraction is thought to be due to the action of fibroblasts remodelling the ECM, to which they are attached through integrin receptors, either by tractional forces as they migrate through the matrix or by contraction after having differentiated into myofibroblasts [106, 124]. The process can be mimicked by cells in collagen gels, where it has been shown that contraction involves the synergistic effects of both α₂β₁ integrin and MMPs [125], the former mediating cell-matrix interactions and the latter facilitating cell migration through the matrix.

Conclusion

First, new imaging techniques are opening up new possibilities for the direct visualisation of proteolytic activity and remodelling both in vitro and in vivo. For example, visualisation of proteolytic cell migration through collagen gels has been shown to lead to re-organisation and alignment of collagen fibres [109]. Similarly, both large scale and local movements of ECM filaments, the latter driven by cell migration, have been observed by time-lapse imaging in developing embryos [126]. In addition, the use of timer reporters has allowed specific molecules to be tracked during morphogenetic changes, as for example during elastin assembly [127]. Finally, both fluorescent and radioactively labelled activity-based probes have been developed to directly visualise sites of proteolytic activity in situ, as opposed to antibody-based methods which usually do not distinguish between precursor and active forms. Such probes are easily designed for serine and cysteine proteases, for which numerous suicide substrates are known, but probes targeting metalloproteases are also under development with interesting results [128].

Also, the discovery of new protease substrates in cell-based assays has gained in efficiency with the development of quantitative proteomics and its application to proteases [1, 66]. Using iTRAQ (isobaric Tags for Relative and Absolute Quantification) and SILAC (Stable Isotope Labelling by Amino-acids in cell Culture) technologies, it is now possible, in one single experiment, to identify a large proportion of the proteins (and their fragments) found in a defined proteome and to quantify their relative abundances in several biological contexts (typically in the presence and absence of a protease).

Together with knock-out mice, these new developments should contribute towards establishing the complex network of interactions involved in each particular remodelling event and permit the use of this knowledge to devise optimized therapeutic strategies.

Acknowledgements

References

2 Doan N, Gettins PG. Human alpha2-macroglobulin is composed of multiple domains, as predicted by homology with complement component C3. Biochem J 2007; 407: 23-30.

3 Lopez-Otin C, Bond JS. Proteases: Multifunctional enzymes in life and disease. J Biol Chem 2008; 283: 30433-7.

4 Lund LR, Romer J, Bugge TH, et al. Functional overlap between two classes of matrix-degrading proteases in wound healing. EMBO J 1999; 18: 4645-56.

5 Lund LR, Green KA, Stoop AA, et al. Plasminogen activation independent of uPA and tPA maintains wound healing in gene-deficient mice. EMBO J 2006; 25: 2686-97.

6 Borgono CA, Diamandis EP. The emerging roles of human tissue kallikreins in cancer. Nat Rev Cancer 2004; 4: 876-90.

7 Szabo R, Bugge TH. Type II transmembrane serine proteases in development and disease. Int J Biochem Cell Biol 2008; 40: 1297-316.

8 Conway K, Ruge F, Price P, et al. Hepatocyte growth factor regulation: an integral part of why wounds become chronic. Wound Repair Regen 2007; 15: 683-92.

9 Lungarella G, Cavarra E, Lucattelli M, et al. The dual role of neutrophil elastase in lung destruction and repair. Int J Biochem Cell Biol 2008; 40: 1287-96.

10 Korkmaz B, Moreau T, Gauthier F. Neutrophil elastase, proteinase 3 and cathepsin G: physicochemical properties, activity and physiopathological functions. Biochimie 2008; 90: 227-42.

11 Garnero P, Borel O, Byrjalsen I, et al. The collagenolytic activity of cathepsin K is unique among mammalian proteinases. J Biol Chem 1998; 273: 32347-52.

12 Yasuda Y, Li Z, Greenbaum D, et al. Cathepsin V, a novel and potent elastolytic activity expressed in activated macrophages. J Biol Chem 2004; 279: 36761-70.

13 Cesarman-Maus G, Hajjar KA. Molecular mechanisms of fibrinolysis. Br J Haematol 2005; 129: 307-21.

14 Grainger DJ, Wakefield L, Bethell HW, et al. Release and activation of platelet latent TGF-beta in blood clots during dissolution with plasmin. Nat Med 1995; 1: 932-7.

15 Sato Y, Tsuboi R, Lyons R, et al. Characterization of the activation of latent TGF-beta by co-cultures of endothelial cells and pericytes or 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 and its functional consequences: role of plasmin and tissue-type plasminogen activator. J Cell Biol 1998; 141: 255-65.

17 Bonnefoy A, Legrand C. Proteolysis of subendothelial adhesive glycoproteins (fibronectin, thrombospondin, and von Willebrand factor) by plasmin, leukocyte cathepsin G, and elastase. Thromb Res 2000; 98: 323-32.

18 Li WY, Chong SS, Huang EY, et al. Plasminogen activator/plasmin system: a major player in wound healing? Wound Repair Regen 2003; 11: 239-47.

19 Ossowski L. guirre-Ghiso JA. Urokinase receptor and integrin partnership: coordination of signaling for cell adhesion, migration and growth. Curr Opin Cell Biol 2000; 12: 613-20.

20 Green KA, Lund LR. ECM degrading proteases and tissue remodelling in the mammary gland. Bioessays 2005; 27: 894-903.

21 Sterchi EE, Stocker W, Bond JS. Meprins, membrane-bound and secreted 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, adamalysins, serralysins, and matrixins (collagenases) define a superfamily of zinc-peptidases. Protein Sci 1995; 4: 823-40.

23 Saarialho-Kere UK, Crouch EC, Parks WC. Matrix metalloproteinase matrilysin is constitutively expressed in adult human exocrine epithelium. J Invest Dermatol 1995; 105: 190-6.

24 Ii M, Yamamoto H, Adachi Y, et al. Role of matrix metalloproteinase-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 metalloproteinases and the regulation of tissue remodelling. Nat Rev Mol Cell Biol 2007; 8: 221-33.

26 Lee MH, Murphy G. Matrix metalloproteinases at a glance. J Cell Sci 2004; 117: 4015-6.

27 Mott JD, Werb Z. Regulation of matrix biology by matrix metalloproteinases. Curr Opin Cell Biol 2004; 16: 558-64.

28 Newby AC. Matrix metalloproteinases regulate migration, proliferation, and death of vascular smooth muscle cells by degrading matrix and non-matrix substrates. Cardiovasc Res 2006; 69: 614-24.

29 Nagase H, Visse R, Murphy G. Structure and function of matrix metalloproteinases and TIMPs. Cardiovasc Res 2006; 69: 562-73.

30 Porter S, Clark IM, Kevorkian L, et al. The ADAMTS metalloproteinases. Biochem J 2005; 386: 15-27.

31 Lee NV, Sato M, Annis DS, et al. ADAMTS1 mediates the release of antiangiogenic polypeptides from TSP1 and 2. EMBO J 2006; 25: 5270-83.

32 Amour A, Slocombe PM, Webster A, et al. TNF-alpha converting enzyme (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 of ADAM-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 potent inhibitor 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 development; a pericellular protein with a homology to the ADAMTS metalloproteinases. Development 2000; 127: 5475-85.

36 Lee NV, Rodriguez-Manzaneque JC, Thai SN, et al. Fibulin-1 acts as a cofactor for the matrix metalloprotease ADAMTS-1. J Biol Chem 2005; 280: 34796-804.

37 Tousseyn T, Jorissen E, Reiss K, et al. (Make) stick and cut loose - disintegrin metalloproteases in development and disease. Birth Defects Res C Embryo Today 2006; 78: 24-46.

38 Edwards DR, Handsley MM, Pennington CJ. The ADAM metalloproteinases. Mol Aspects Med 2008; 29: 258-89.

39 Hopkins DR, Keles S, Greenspan DS. The bone morphogenetic protein 1/Tolloid-like metalloproteinases. Matrix Biol 2007; 26: 508-23.

40 Ge G, Greenspan DS. BMP1 controls TGFbeta1 activation via cleavage of latent TGFbeta-binding protein. J Cell Biol 2006; 175: 111-20.

41 Amano S, Scott IC, Takahara K, et al. Bone morphogenetic protein 1 is an extracellular processing enzyme of the laminin 5 gamma2 chain. J Biol Chem 2000; 275: 22728-35.

42 Rattenholl A, Pappano WN, Koch M, et al. Proteinases of the bone morphogenetic protein-1 family convert procollagen VII to mature anchoring fibril collagen. J Biol Chem 2002; 277: 26372-8.

43 Gonzalez EM, Reed CC, Bix G, et al. BMP-1/Tolloid-like metalloproteases process endorepellin, the angiostatic C-terminal fragment of perlecan. J Biol Chem 2005; 280: 7080-7.

44 Ge G, Fernandez CA, Moses MA, et al. Bone morphogenetic protein 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 tolloid proteinases. Cell 2006; 124: 147-59.

46 Kobayashi K, Luo M, Zhang Y, et al. Secreted Frizzled-related protein 2 is a procollagen C proteinase enhancer with a role in fibrosis associated with myocardial infarction. Nat Cell Biol 2009; 11: 46-55.

47 Moali C, Font B, Ruggiero F, et al. Substrate-specific modulation of a multisubstrate proteinase. C-terminal processing of fibrillar procollagens is the only BMP-1-dependent activity to be enhanced by PCPE-1. J Biol Chem 2005; 280: 24188-94.

48 Scott IC, Blitz IL, Pappano WN, et al. Homologues of Twisted gastrulation are extracellular cofactors in antagonism of BMP signalling. Nature 2001; 410: 475-8.

49 Inomata H, Haraguchi T, Sasai Y. Robust stability of the embryonic axial pattern requires a secreted scaffold for chordin degradation. Cell 2008; 134: 854-65.

50 Chen BK, Leiferman KM, Pittelkow MR, et al. Localization and regulation of pregnancy-associated plasma protein a expression in healing human skin. J Clin Endocrinol Metab 2003; 88: 4465-71.

51 Glerup S, Boldt HB, Overgaard MT, et al. Proteinase inhibition by proform of eosinophil major basic protein (pro-MBP) is a multistep process of intra- and intermolecular disulfide rearrangements. J Biol Chem 2005; 280: 9823-32.

52 Gyrup C, Oxvig C. Quantitative analysis of insulin-like growth factor-modulated proteolysis of insulin-like growth factor binding protein-4 and -5 by pregnancy-associated plasma protein-A. Biochemistry 2007; 46: 1972-80.

53 Hirano M, Ma BY, Kawasaki N, et al. Mannan-binding protein blocks the activation of metalloproteases meprin alpha and beta. J Immunol 2005; 175: 3177-85.

54 Singer AJ, Clark RA. Cutaneous wound healing. N Engl J Med 1999; 341: 738-46.

55 Toriseva M, Kahari VM. Proteinases in cutaneous wound healing. 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 factors 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 shedding from platelets in vitro and in vivo. Circ Res 2004; 95: 677-83.

59 Tsirogianni AK, Moutsopoulos NM, Moutsopoulos HM. Wound healing: immunological aspects. Injury 2006; 37 (Suppl 1): S5-S12.

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. Increased inflammation 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-8 facilitates neutrophil migration through the corneal stromal matrix by collagen 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 metalloproteinases as modulators of inflammation and innate immunity. Nat Rev Immunol 2004; 4: 617-29.

64 Wolf M, Albrecht S, Marki C. Proteolytic processing of chemokines: implications in physiological and pathological conditions. Int J Biochem Cell Biol 2008; 40: 1185-98.

65 McQuibban GA, Gong JH, Wong JP, et al. Matrix metalloproteinase processing of monocyte chemoattractant proteins generates CC chemokine receptor antagonists with anti-inflammatory properties in vivo. Blood 2002; 100: 1160-7.

66 Dean RA, Overall CM. Proteomics discovery of metalloproteinase substrates in the cellular context by iTRAQ labeling reveals a diverse MMP-2 substrate degradome. Mol Cell Proteomics 2007; 6: 611-23.

67 Ishida Y, Gao JL, Murphy PM. Chemokine receptor CX3CR1 mediates skin wound healing by promoting macrophage and fibroblast accumulation and function. J Immunol 2008; 180: 569-79.

68 Yu Q, Stamenkovic I. Cell surface-localized matrix metalloproteinase-9 proteolytically activates TGF-beta and promotes tumor 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 of all trades” proteoglycan of cartilaginous weight-bearing connective tissues. Bioessays 2008; 30: 457-69.

71 Bergers G, Brekken R, McMahon G, et al. Matrix metalloproteinase-9 triggers the angiogenic switch during carcinogenesis. Nat Cell Biol 2000; 2: 737-44.

72 Roy R, Zhang B, Moses MA. Making the cut: protease-mediated regulation of angiogenesis. Exp Cell Res 2006; 312: 608-22.

73 Lee S, Jilani SM, Nikolova GV, et al. Processing of VEGF-A by matrix metalloproteinases regulates bioavailability and vascular patterning in tumors. J Cell Biol 2005; 169: 681-91.

74 Dean RA, Butler GS, Hamma-Kourbali Y, et al. Identification of candidate angiogenic inhibitors processed by matrix metalloproteinase 2 (MMP-2) in cell-based proteomic screens: disruption of vascular endothelial growth factor (VEGF)/heparin affin regulatory peptide (pleiotrophin) and VEGF/Connective tissue growth factor angiogenic inhibitory complexes by MMP-2 proteolysis. Mol Cell Biol 2007; 27: 8454-65.

75 Bunn RC, Fowlkes JL. Insulin-like growth factor binding protein proteolysis. Trends Endocrinol Metab 2003; 14: 176-81.

76 Rifkin DB. Latent transforming growth factor-beta (TGF-beta) binding proteins: orchestrators of TGF-beta availability. J Biol Chem 2005; 280: 7409-12.

77 Isogai Z, Ono RN, Ushiro S, et al. Latent transforming growth factor beta-binding protein 1 interacts with fibrillin and is a microfibril-associated protein. J Biol Chem 2003; 278: 2750-7.

78 Neptune ER, Frischmeyer PA, Arking DE, et al. Dysregulation of TGF-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 transforming growth factor beta1 - an intimate relationship. Eur J Cell Biol 2008; 87: 601-15.

80 Mu D, Cambier S, Fjellbirkeland L, et al. The integrin alpha(v)beta8 mediates epithelial homeostasis through MT1-MMP-dependent activation of TGF-beta1. J Cell Biol 2002; 157: 493-507.

81 Schultz-Cherry S, Murphy-Ullrich JE. Thrombospondin causes activation of latent transforming growth factor-beta secreted by endothelial cells by a novel mechanism. J Cell Biol 1993; 122: 923-32.

82 Ribeiro SM, Poczatek M, Schultz-Cherry S, et al. The activation sequence of thrombospondin-1 interacts with the latency-associated peptide 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 contraction activates latent TGF-beta1 from the extracellular matrix. J Cell Biol 2007; 179: 1311-23.

84 Higashiyama S, Nanba D. ADAM-mediated ectodomain shedding 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 of the carboxy-terminal fragment of proHB-EGF causes nuclear export of PLZF. 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 of Notch signaling in wound healing. PLOS One 2009; 2: e1167.

88 Murphy G, Murthy A, Khokha R. Clipping, shedding and RIPping keep immunity on cue. Trends Immunol 2008; 29: 75-82.

89 Chambers RC, Laurent GJ. Coagulation cascade proteases and tissue 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 in mice 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 receptors. Matrix Biol 2004; 23: 75-85.

93 Remy L, Trespeuch C, Bachy S, et al. Matrilysin 1 influences colon carcinoma cell migration by cleavage of the laminin-5 beta3 chain. 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 vascular smooth muscle cell proliferation. Cardiovasc Res 2009; 81: 178-86.

95 Noe V, Fingleton B, Jacobs K, et al. Release of an invasion promoter E-cadherin fragment by matrilysin and stromelysin-1. J Cell Sci 2001; 114: 111-8.

96 Maretzky T, Reiss K, Ludwig A, et al. ADAM10 mediates E-cadherin shedding and regulates epithelial cell-cell adhesion, migration, 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 cleavage of E-cadherin by leukocyte elastase during acute experimental pancreatitis in rats. Gastroenterology 2005; 129: 1251-67.

98 Huguenin M, Muller EJ, Trachsel-Rosmann S, et al. The metalloprotease meprinbeta processes E-cadherin and weakens intercellular adhesion. 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: Implications 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 Biol 2005; 15: 52-60.

102 Stringa E, Knauper V, Murphy G, et al. Collagen degradation and platelet-derived growth factor stimulate the migration of vascular smooth muscle cells. J Cell Sci 2000; 113 (Pt 11): 2055-64.

103 Xu J, Rodriguez D, Petitclerc E, et al. Proteolytic exposure of a cryptic site within collagen type IV is required for angiogenesis and tumor 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-2 in a model of concomitant resistance. J Biol Chem 1999; 274: 29568-71.

105 Mannello F, Luchetti F, Falcieri E, et al. Multiple roles of matrix metalloproteinases during apoptosis. Apoptosis 2005; 10: 19-24.

106 Rhee S, Grinnell F. Fibroblast mechanics in 3D collagen matrices. Adv Drug Deliv Rev 2007; 59: 1299-305.

107 Hinz B. Formation and function of the myofibroblast during tissue repair. J Invest Dermatol 2007; 127: 526-37.

108 Zaman MH, Trapani LM, Sieminski AL, et al. Migration of tumor cells 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-step process. Cancer Metastasis Rev, 2009.

110 Alexander NR, Branch KM, Parekh A, et al. Extracellular matrix rigidity 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 nucleators. Curr Opin Cell Biol 2008; 20: 495-501.

113 Gopalakrishnan B, Wang WM, Greenspan DS. Biosynthetic processing of the Pro-alpha1(V)Pro-alpha2(V)Pro-alpha3(V) procollagen heterotrimer. J Biol Chem 2004; 279: 30904-12.

114 Scott IC, Imamura Y, Pappano WN, et al. Bone morphogenetic protein-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 enhance its ability to regulate collagen fibrillogenesis. J Biol Chem 2004; 279: 41626-33.

116 Lucero HA, Kagan HM. Lysyl oxidase: an oxidative enzyme and effector of cell function. Cell Mol Life Sci 2006; 63: 2304-16.

117 Telci D, Griffin M. Tissue transglutaminase (TG2) - a wound response enzyme. Front Biosci 2006; 11: 867-82.

118 Uzel MI, Scott IC, Babakhaniou-Chase H, et al. Multiple bone morphogenetic protein 1-related mammalian metalloproteinases process pro-lysyl oxidase at the correct physiological site and control lysyl oxidase activation in mouse embryo fibroblast cultures. J Biol Chem 2001; 276: 22537-43.

119 Atsawasuwan P, Mochida Y, Katafuchi M, et al. Lysyl oxidase binds transforming growth factor-beta and regulates its signaling via amine oxidase activity. J Biol Chem 2008; 283: 34229-40.

120 Metcalfe AD, Ferguson MW. Tissue engineering of replacement skin: the crossroads of biomaterials, wound healing, embryonic development, stem cells and regeneration. J R Soc Interface 2007; 4: 413-37.

121 Caplan AI. Embryonic development and the principles of tissue engineering. Novartis Found Symp 2003; 249: 17-25.

122 Shah M, Foreman DM, Ferguson MWJ. Neutralising antibody to TGF-b1,2 reduces cutaneous scarring in adult rodents. J Cell Sci 1994; 107: 1137-57.

123 Connon CJ, Meek KM. Organization of corneal collagen fibrils during the healing of trephined wounds in rabbits. Wound Repair Regen 2003; 11: 71-8.

124 Tomasek JJ, Gabbiani G, Hinz B, et al. Myofibroblasts and mechano-regulation of connective tissue remodelling. Nat Rev Mol Cell Biol 2002; 3: 349-63.

125 Phillips JA, Bonassar LJ. Matrix metalloproteinase activity synergizes 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 macroassembly dynamics in early vertebrate embryos. Curr Top Dev Biol 2006; 73: 237-58.

127 Kozel BA, Rongish BJ, Czirok A, et al. Elastic fiber formation: a dynamic view of extracellular matrix assembly using timer reporters. J Cell Physiol 2006; 207: 87-96.

128 Blum G. Use of fluorescent imaging to investigate pathological protease activity. Curr Opin Drug Discov Devel 2008; 11: 708-16.