Conformational Dependence of Collagenase (Matrix Metalloproteinase-1) Up-regulation by Elastin Peptides in Cultured Fibroblasts*

We have established that treatment of cultured human skin fibroblasts with tropoelastin or with heterogenic peptides, obtained after organo-alkaline or leukocyte elastase hydrolysis of insoluble elastin, induces a high expression of pro-collagenase-1 (pro-matrix metalloproteinase-1 (pro-MMP-1)). The identical effect was achieved after stimulation with a VGVAPG synthetic peptide, reflecting the elastin-derived domain known to bind to the 67-kDa elastin-binding protein. This clearly indicated involvement of this receptor in the described phenomenon. This notion was further reinforced by the fact that elastin peptides-dependent MMP-1 up-regulation has not been demonstrated in cultures preincubated with 1 mm lactose, which causes shedding of the elastin-binding protein and with pertussis toxin, which blocks the elastin-binding protein-dependent signaling pathway involving G protein, phospholipase C, and protein kinase C. Moreover, we demonstrated that diverse peptides maintaining GXXPG sequences can also induce similar cellular effects as a “principal” VGVAPG ligand of the elastin receptor. Results of our biophysical studies suggest that this peculiar consensus sequence stabilizes a type VIII β-turn in several similar, but not identical, peptides that maintain a sufficient conformation to be recognized by the elastin receptor. We have also established that GXXPG elastin-derived peptides, in addition to pro-MMP-1, cause up-regulation of pro-matrix metalloproteinase-3 (pro-stromelysin 1). Furthermore, we found that the presence of plasmin in the culture medium activated these MMP proenzymes, leading to a consequent degradation of collagen substrate. Our results may be, therefore, relevant to pathobiology of inflammation, in which elastin-derived peptides bearing the GXXPG conformation (created after leukocyte-dependent proteolysis) bind to the elastin receptor of local fibroblasts and trigger signals leading to expression and activation of MMP-1 and MMP-3, which in turn exacerbate local connective tissue damage.

The extracellular matrix protein elastin is responsible for the elastic properties of tissues such as lung, skin, and large arteries (1)(2)(3). Due to its numerous cross-links and the extreme hydrophobicity of its tropoelastin chains, elastin is highly resistant to proteolysis. However, during inflammatory disorders, proteinases secreted from polymorphonuclear neutrophils, such as elastase, cathepsin G, and gelatinase B may cause significant elastolysis (4).
Matrix metalloproteinases (MMPs) are potent proteinases involved in a broad range of normal and pathological processes (25). Their expression is regulated through interaction between cells and extracellular matrix via several classes of cell surface receptors (25). For example, a single module within fibronectin can elicit a particular cell response, resulting in either up-or down-regulation of collagenase (MMP-1) production (26). Attracted by the wide range of cellular effects induced by elastinderived peptides, we investigated whether peptides resulting from elastin degradation would also control MMPs (MMP-1, i.e. collagenase-1; MMP-3, i.e. stromelysin-1) expression and secretion and thereby contribute to further degradation of other matrix components.
Results of the present study indicate that cultured human skin fibroblasts up-regulate expression and secretion of pro-MMP-1 and pro-MMP-3 after stimulation with products of organo-alkaline-and human leukocyte elastase (HLE)-dependent elastin degradation. Our data suggest that among many elas-* This work was supported by grants from the Association Régionale pour l'Enseignement et la Recherche Scientifique et Technologique (ARERS) and the Région Champagne-Ardenne (to W. H.), by an Association pour la Recherche sur le Cancer (ARC) fellowship (to B. B.), by a grant from the Ligue contre le Cancer (to H. E.), and by CNRS (UPRES-A 6021). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
* tin-derived peptides, only those with the GXXPG consensus sequence possess a conformation that allows binding to the EBP and a consequent triggering of signals responsible for the up-regulation pro-MMP-1 and pro-MMP-3. Finally, we have established that both secreted pro-enzymes can be activated after the addition of exogenous plasmin to the culture medium and then degrade a collagen substrate. Our results thus raise the possibility that elastin degradation could lead to collagenolysis during normal and/or pathological conditions.

EXPERIMENTAL PROCEDURES
Samples Preparation and Reagents-Bovine tropoelastin was obtained from Elastin Products Co. (Universal Biological Ltd., London).
-Elastin was prepared from insoluble bovine elastin purified from calf ligamentum nuchae using the hot alkali procedure (27).
Alternatively, insoluble elastin was partially hydrolyzed by HLE (Universal Biological Ltd., London). Briefly, 75 mg of elastin ground to Ͻ100 mesh were dispersed in 7.5 ml of 100 mM Tris, 5 mM CaCl 2 , pH 8.0, and maintained under constant mechanical stirring for 1 h at 37°C. Human leukocyte elastase (37.5 g) was added, and the flasks were incubated for 1 h or 24 h at 37°C. The reaction was stopped by the addition of 1 mM phenylmethanesulfonyl fluoride. Elastin dispersions were then centrifuged at 12,000 ϫ g for 15 min. Supernatants were removed, and pellets were washed three times with 10 ml of distilled water, lyophilized, and weighed. Due to the low Tyr content of elastin, the protein content of supernatants (hydrolysates) was determined (28) using -elastin as standard rather than bovine serum albumin.
Peptides Synthesis-The synthetic peptides VGVAPG, GVAPGV, VAPGVG, APGVGV, PGVGVA, GVGVAP, PGAIPG, and LGTIPG were purchased from Ansynth Service B.V. (Roosendaal, The Netherlands) or synthesized according to classical solid phase synthesis (10,22). Purity of the peptides was confirmed by high performance liquid chromatography and by fast atom bombardment mass spectrometry.
Cell Culture-Human skin fibroblast strains were established from explants of human adult skin biopsies obtained from informed healthy volunteers (age 21-41 years). Cells were grown as monolayer cultures in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and 2 mM glutamine in the presence of 5% CO 2 . Cells at subcultures 5 to 10 were used. Fibroblasts were grown to subconfluency in 10% serum containing medium. Cell cultures were washed twice with phosphate-buffered saline and incubated for 6 or 24 h in serum-free medium with or without elastin-like material in the presence or absence of lactose and different cell-signaling effectors. The culture supernatant was then harvested, and cellular debris were pelleted (500 ϫ g, 10 min, 22°C).
Collagen Degradation-Petri dishes were coated with 150 g of 3 Hradiolabeled type I collagen from rat tail tendon in 18 mM acetic acid (24000 cpm/100 g collagen). Human skin fibroblasts were grown on this collagen matrix in a medium supplemented with 1% fetal calf serum in the presence or absence of kE (50 g/ml) and/or plasmin (0.48 units/ml). After 24 h, the culture supernatant was harvested, and its radioactivity was measured.
Western Blot Analysis-After medium concentration, the individual fractions were adjusted to the same protein concentration, electrophoresed in a 0.1% SDS-10% polyacrylamide gel under reducing conditions, and transferred onto Immobilon-P membranes (Millipore, Saint-Quentin-en-Yvelines, France). The membranes were saturated with 5% calf serum, 0.1% Tween 20 in Tris-buffered saline for 2 h, incubated for 1 h with sheep polyclonal anti-human MMP-1 or rabbit polyclonal antihuman MMP-3 antibodies, and then incubated with the alkaline phosphatase-conjugated anti-sheep or anti-rabbit antibodies for 1 h at room temperature. Immuno complexes were visualized with nitro blue tetrazolium 5-bromo-4-chloro-3-indolyl phosphate reagent. The molecular masses corresponding to the stained bands were determined, and they were further quantified by densitometry with the Bio1D software (Vilber-Lourmat, Marne-la-Vallée, France). Linear range of intensity of the bands was assessed using purified pro-MMP-1 or pro-MMP-3 as standard. Linearity was between 10 and 200 ng of enzyme.
Northern Blot Analysis-Confluent cultures were washed twice with serum-free medium and then incubated under serum-free conditions for 2 h. Subsequently, elastin-like material was added to the medium, and incubation was continued for 6 or 24 h. The cultures were then washed with phosphate-buffered saline, and total RNA was extracted from fibroblasts as described (29). After washing, the blots were exposed to Kodak X-Omat film at Ϫ80°C using intensifying screens. The human 2.1-kilobase MMP-1 cDNA and 1.9-kilobase MMP-3 cDNA probes were kind gifts from Dr. Angel (Deutsches Krebsforschungs-zentrum, Heidelberg, Germany) and Dr. Saus (Valencia Foundation of Biomedical Investigations, Valencia, Spain), respectively. The radioactive bands were quantified by densitometry and normalized using the human 1.06-kilobase 36B4 cDNA probe generously provided by Prof. Chambon (University of Strasbourg, Strasbourg, France). 36B4 is a reporter gene encoding the human acidic ribosomal phosphoprotein PO (30).
CD Spectroscopy-The CD spectra were recorded in 0.1-cm path length cylindrical cells on a JASCO J-810 dichrograph by averaging three consecutive scans. The samples were dissolved in water at concentrations ranging from 1 to 3 ϫ 10 Ϫ3 M. The data are presented in terms of mean residue molar ellipticity expressed in deg cm 2 dmol Ϫ1 in the 185-250-nm spectral range.
Structural Predictions-The ␤-turns propensities (P t ) have been calculated using our software COUDES (TURNS). 2 They represented the propensity (31) for a tetrapeptide to belong to a particular type of ␤-turn. If P t Ͼ 1, the tetrapeptide is considered as a probable turn; if P t Ͻ 1, the tetrapeptide is not taken as a turn. The calculation of P t is based on the simple product of four individual residue propensities. These residue propensities, defined for each type of turn and taking into account the location of the residue in the turn (i.e. the position 1, 2, 3, or 4), were determined from a reference set of 205 known three-dimensional structures of proteins (32).
Statistical Analysis-Experiments were performed in triplicate. Results are expressed as means Ϯ S.E. Differences between control means and treated groups were assessed using the unpaired Student's t test. (33), human skin fibroblasts in culture produced a low level of pro-MMP-1 ( Fig. 1; Control). The two immuno-reactive species had an apparent molecular mass of 57 and 53 kDa and corresponded to the glycosylated and nonglycosylated pro-MMP-1 isoforms, respectively. We found that the level of pro-MMP-1 was considerably enhanced when cells were incubated in the presence of 50 g/ml elastin for 24 h ( Fig. 1; Elastin). Incubation of fibroblasts, with the supernatant withdrawn after a 24-h hydrolysis of elastin by elastase (50 g of elastin-derived 2 P. Fuchs and A. J. P. Alix, article in preparation. peptides/ml), resulted in a 2-fold stronger stimulation ( Fig. 1; HLE lysate). The corresponding insoluble elastin pellet had lost this potentiality ( Fig. 1; HLE pellet), suggesting that the elastin peptides promoting pro-MMP-1 production were totally released from insoluble elastin by HLE.

Elastin and Elastin Peptides Up-regulate Pro-MMP-1 Production and Expression by Fibroblasts-As previously reported
Interestingly, tropoelastin, kE, and the VGVAPG synthetic peptide could also stimulate pro-MMP-1 production (Fig. 1). The fact that tropoelastin also exhibited an effect excluded the possibility that elastin-cross-linked regions could be involved. Consequently, the strong stimulation observed after incubation with kE ( Fig. 1), a mixture of elastin-derived peptides of a diverse length, particularly rich in fragments originating from the hydrophobic domains of tropoelastin, was explained. As VGVAPG proved an efficient stimulator of pro-MMP-1 production ( Fig. 1), we hypothesized that this sequence and similar ones could be responsible for pro-MMP-1 up-regulation.
The kE-induced pro-MMP-1 production stimulation was correlated with enhanced expression of MMP-1 mRNA levels ( Fig.  2), suggesting that elastin peptides up-regulated MMP-1 at the expression level. Standardization of data using a 36B4 cDNA probe (Fig. 2) demonstrated that, after 24 h of culture, MMP-1 mRNA levels were increased 8-fold.
We point out here that although elastin, elastolysate, tropoelastin, and kE concentrations as low as 50 g/ml proved sufficient to stimulate pro-MMP-1 production, comparable stimulation levels could only be reached with 200 g/ml VGVAPG, a major ligand domain of the elastin receptor. Although this concentration was 3-4 orders of magnitude higher than the one required for some other elastin peptide-induced activities such as chemotaxis, the used concentration range was similar with those needed for enzyme excretion (34) and proliferation (15).
The 67-kDa EBP Mediates the Effect of Elastin Peptides on Pro-MMP-1 Production by Fibroblasts-Elastin peptides bearing the VGVAPG sequence have been shown as a principal ligand of the 67-kDa EBP (35). It has also been established that the EBP interaction with this elastin-derived domain was only possible in the absence of galactosugars, which otherwise may bind to a separate galactolectin binding domain of the EBP and make this molecule unreceptive for elastin. Thus, the addition of such galactosugar-bearing moieties as lactose blocks the specific interaction between elastin peptides and the EBP (5).
Indeed, the addition of 1 mM lactose to the fibroblast culture medium resulted in a substantial (35%) inhibition of kE-stimulated pro-MMP-1 production (Fig. 3). In the same conditions, VGVAPG-stimulating effect was inhibited by 80% (Fig. 3). These data strongly suggested that binding of VGVAPG on the 67-kDa EBP could explain pro-MMP-1 up-regulation. It needs to be emphasized, however, that stimulation of pro-MMP-1 by interleukin-1␤ could not be blocked by lactose, and lactose alone had no effect on pro-MMP-1 accumulation (data not shown).
An additional experiment aiming at the elucidation of the signaling pathways triggered upon interaction between elastin peptides and EBP has also been carried out. Several inhibitors of EBP-dependent intracellular signaling have been tested. The results listed in Table I clearly indicate the involvement of a pertussis toxin-sensitive G protein, phospholipase C, and protein kinase C (but not phospholipase D and protein-tyrosine kinase) in the intracellular signaling pathways leading to pro-MMP-1 up-regulation after exposure to elastin-derived ligands. These results further implicate involvement of the EBP in the signaling pathways leading to up-regulation of MMP-1 and are consistent with the previously described EBP-dependent signaling during elastin peptide-stimulated chemotaxis of leukocytes (17). It must also be stressed that elastin peptide-dependent induction of pro-MMP-1 could not be blocked by an interleukin-1 receptor antagonist (Table I).
Peptides Containing the GXXPG Consensus Sequence Upregulate Pro-MMP-1-The multiple hydrophobic VGVAPG sequences occur exclusively in tropoelastin region encoded by exon 24 (36). In bovine tropoelastin, it repeats twice, and in human tropoelastin, it repeats six times (37). Since the synthetic peptide reflecting VGVAPG sequence proved so efficient in stimulation of pro-MMP-1 production (Fig. 1), we also tested whether other domains bearing a similar conformation could evoke similar cellular effects.
Peptides corresponding to circular permutation of the

FIG. 2. Influence of kE on MMP-1 mRNA levels in human skin fibroblasts.
Northern blot analysis of 20 g of total RNA from unstimulated fibroblasts at 6 h (control 6h) or 24 h (control 24h) and treated with 50 g of kE/ml for 6 h (kE 6h) or 24 h (kE 24h). Histograms of the ratio MMP-1 mRNA versus mRNA from the constitutive 36B4 gene is provided. Statistically significant differences between stimulated and control are indicated (*, p Ͻ 0.01).

FIG. 3. Western blot analysis with anti-pro-MMP-1 antibody shows that 1 mM lactose inhibits the kE-and VGVAPG-induced production of pro-MMP-1 in a 24-h culture of human skin fibroblast production after a 24-h incubation. -Elastin concentration was 50 g/ml.
VGVAPG sequence (VGVAPG, GVAPGV, VAPGVG, APGVGV, PGVGVA, GVGVAP) were used to identify the precise sequences and/or motifs involved in pro-MMP-1 up-regulation. Two other 67-kDa EBP binding peptides were also used for comparison: PGAIPG and LGTIPG, from elastin (9) and laminin B1 chain (35), respectively. As presented in Fig. 4, only those peptides bearing the GXXPG sequence could induce pro-MMP-1, suggesting that this consensus motif was important for correct binding to the EBP. VGVAPG induced pro-MMP-1 to a substantial level (1 ng/h/10 5 fibroblasts), and GVAPGV was 20% more efficient. The MMP-1 stimulation efficiency was GVAPGV Ͼ VGVAPG Ͼ LGTIPG Ͼ PGAIPG, suggesting that the nature of the residues found at X positions in GXXPG could determine the affinity of the peptide pattern for binding to EBP. Apolar side chains seemed to be preferred but not required, as demonstrated by the presence of a Thr residue in LGTIPG.
Pro-MMP-1 Up-regulation by EDPs Depends on the Peptide Conformation-To understand the importance of the GXXPG sequence for EBP binding, the conformation of the peptides was investigated using CD spectroscopy. Our CD spectra were in excellent agreement with those obtained by others for similar peptides (12,38). They were characterized by a dominant negative band centered around 200 nm, commonly associated with the -* electronic transition of disordered peptides (Fig.  5). Nevertheless, two spectral groups could be defined: one in which the negative band was centered around 195 nm (filled symbols) and another exhibiting a band around 200 nm (open symbols). Strikingly, the spectra of peptides that did not enhance pro-MMP-1 production all belonged to the first group, whereas those of active peptides were in the second one. This separation was underlined by the consistency of group 1 minimal intensities, about Ϫ9000 deg cm 2 dmol Ϫ1 (Fig. 5) as compared with those of the second group. These results suggested that the intensity of pro-MMP-1 production to the tested peptides could be conformational-dependent, although those peptides were substantially unordered.
Indeed, two groups could be distinguished within the spectra of active peptides: LGTIPG and PGAIPG exhibited a negative maximum around 196 nm and a shoulder at 220 nm, whereas those of VGVAPG and GVAPGV yielded a highly symmetric minimum around 200 nm. Interestingly the most active peptides were from the second group. These findings suggested that activity of the peptides originated from the presence of several folded peptides within a population of unordered conformers.
Our prediction data (Table II) suggested that all the active peptides contained a type VIII ␤-turn (39) in the GXXP se-quence (Fig. 6). The glycyl residue after the GXXP turn was also necessary for the biological activity, since GVGVAP was inactive even if it comprised the GXXP sequence.
Although a type VIII ␤-turn might be adopted by the active peptides, it must be stressed that the peptides consist of only six residues and were thus very flexible. Therefore, they could form a type VIII ␤-turn but probably not as their dominant conformation. The fact that a type VIII ␤-turn together with the PG sequence was a common feature of the biologically active hexapeptides indicates that this conformation could well constitute the dominant structure needed for the binding to their target, i.e. cell surface EBP.
The CD spectra of peptides could not be interpreted in terms of the possible presence of type VIII ␤-turns because, to our knowledge, CD spectra concerning this particular class of ␤-turns were not available. However, considering that the essential dihedrals (i.e. those pertaining to residues iϩ1 and iϩ2, respectively) were in the right-handed helical and ␤ regions of a Ramachadran plot, one might suggest that the type VIII ␤-turn should give dominant negative contribution to the CD spectrum at wavelengths longer than 190 nm. In the case of the active peptides, their CD spectra were fully compatible with the presence of type VIII ␤-turn and open (unordered) conformations.
Elastin Peptides Up-regulate Pro-MMP-3, a Potential Activator of Pro-MMP-1-To further explore the physiopathological significance of pro-MMP-1 stimulation by elastin peptides, we  4. Influence of elastin-and laminin-derived peptides on the production of pro-MMP-1 by human skin fibroblasts. Western blots analyses of pro-MMP-1 production after a 24-h incubation. All synthetic peptides were used at 200 g/ml. Statistically significant differences between stimulated and control are indicated (*, p Ͻ 0.01; **, p Ͻ 0.001). investigated whether the EBP-dependent cellular response would also involve activation of up-regulated pro-MMP-1 to MMP-1. Since pathways of pro-MMP-1 activation involve MMP-3 and/or the plasmin system (40), we also tested these possibilities.
Western blotting indicated that fibroblasts stimulated either with kE or with VGVAPG significantly up-regulated expression of proteins reacting with anti-MMP-3 antibody. Two immuno-reactive bands corresponding to pro-MMP-3 glycosylated (60 kDa) and nonglycosylated (57 kDa) isoforms were observed and the accumulation of pro-MMP-3 in the medium was decreased when cells were treated with 1 mM lactose (Fig. 7). Furthermore the appearance of these pro-MMP-3 bands was consistent with an increased MMP-3 gene expression in kEstimulated fibroblasts (Fig. 8). The demonstrated up-regulation of MMP-3 expression was strikingly parallel to that observed for MMP-1, suggesting that both MMP-1 and MMP-3 genes were coregulated by elastin peptides.
The Up-regulation of Pro-MMPs by Elastin Peptides Can Lead to Collagenolysis-To further justify the pathophysiolog-ical relevance of the observed overexpression of both pro-MMPs in cells stimulated by the elastin-derived peptides, the involvement of plasmin was additionally tested. Fibroblasts were cultured on a layer of radiolabeled type I collagen in the presence or absence of kE and/or plasmin to detect collagenolysis. Activation of pro-MMP-1 to MMP-1 was simultaneously monitored by detection of the active form of the enzyme by Western blots and by its ability to degrade radiolabeled collagen (Fig. 9).
Our data indicate that the addition of plasmin to the culture media lead to activation of all secreted pro-MMP-1 to MMP-1 (48 and 42 kDa) in both untreated and kE-stimulated fibroblasts. A similar activation pattern was observed for MMP-3 (data not shown). Apparently, in cultures of unstimulated fi-

APGV
1.03 52.95 --- LGTIPG 1 LGTI 6. The type VIII ␤-turn conformation proposed for the active peptides, as applied to the sequence GVAP. C ␣ carbons are shown as black spheres, and the backbone bending is depicted as a gray ribbon.
broblasts, the basic level of detected MMPs was not sufficient to up-regulate a basic level of collagenolysis. In kE-stimulated fibroblasts, the addition of plasmin triggered a massive activation of up-regulated pro-MMP-1 to a collagenolytic enzyme (Fig. 9). DISCUSSION It has been shown previously that interaction of elastinderived peptides with the cell surface EBP leads to modulation of diverse gene expression and multiple cellular effects (17,20). Our data demonstrate that tropoelastin and elastin degradation products are potent inducers of collagenolytic enzyme expression in human skin fibroblasts. Interestingly, such an effect was attained even at elastin-derived peptide concentrations close to those determined in physiological fluids (41) and potentiated at higher concentrations, often detected during inflammatory processes (33). The fact that this effect could be largely inhibited in the presence of lactose and reproduced by stimulation with VGVAPG and other peptides bearing the GXXPG consensus sequence suggests involvement of the EBP in signaling triggering pro-MMP-1 and pro-MMP-3 up-regulation.
Among elastin-derived peptides, only those containing the GXXPG conformation have been identified as the ligands of the cell surface EBP and stimulators of numerous cellular effects (9). Our present results broaden the range of the biological activity of those peculiar peptides and provides a molecular explanation of their binding to the cell surface receptor.
The assembly of tropoelastin into mature elastic fibers is also directed by the EBP (42). Therefore, it could be assumed that neither tropoelastin bound to this chaperone protein during its secretion from cells nor insoluble elastin surrounded in tissues by a mantle of microfibrils (43) could interact directly with cells under physiological conditions.
The VGVAPG cell recognition domains are accessible on the surface of growing elastic fibers as shown using specific monoclonal antibodies (44). In mature fibers, however, these hydrophobic sequences remain probably masked and simply contribute to the global elasticity of the polymer. However, after tissue injury and release of potent elastolytic enzymes by leukocytes, these domains could be unmasked and/or cleaved of the elastin polymer so that they could bind to the EBP of the adjacent cells.
In fact, we demonstrated that peptides capable of interaction with the EBP can be released from the elastin network by HLE. We therefore suggest that these peptides bound to the EBP and triggered a pathway of intracellular signals leading to the described up-regulation of pro-MMP-1 and pro-MMP-3 production.
Matrix proteins like elastin, laminins, collagens, fibrillins, or fibronectin contain several GXXPG consensus sequences. As seen here, the stimulation of MMP-1 expression could also be achieved using the laminin-derived LGTIPG peptide. Therefore, it is reasonable to assume that small peptides bearing GXXPG sequences, released from other matrix proteins could also interact with the cell surface EBP. We therefore propose that peptides bearing GXXPG conformation, regardless their origin, could serve as stimulators of pro-MMP-1 production once released. This seems to be particularly relevant to inflammatory processes in which HLE and other proteinases released by infiltrating leukocytes could degrade elastin and other matrix proteins, leading to accumulation of peptides, which in turn could interact with EBP and lead to the local accumulation of pro-MMP-1 and its activator pro-MMP-3. Since these pro-enzymes could be activated to active proteinases by plasmin (40,45), degradation of local collagen could follow. In summary, we suggest that initial elastolysis could lead to a consequent degradation of collagen and other matrix compo-nents. This phenomenon could play an important part in the mechanisms controlling connective tissue remodeling during normal and/or pathological processes.