| Sign In to gain access to subscriptions and/or personal tools. |
Elastolysis Induces Collagenolysis in a Gingival Lamina Propria Model
1 Laboratoire Interface Biomatériaux/Tissus Hôtes, INSERM ERM 0203, Institut "Biomolécules" (IFR53), Faculté d’Odontologie, Université de Reims Champagne-Ardenne, 1 rue Maréchal Juin, 51095 Reims Cedex, France; and Correspondence: * corresponding author, 2 rue du Général Koenig, 51100 Reims, France; sandrine.lorimier{at}univ-reims.fr
Elastin peptides were previously reported to increase MMP expression in several cell types. We found binding of these peptides to their receptors led to enhanced MMP-3 and MMP-1 expression, but not activation, in human gingival fibroblasts cultured on plastic dishes. We hypothesized that these peptides, in a more physiological environment, might additionally trigger an MMP-3/MMP-1 activation cascade, leading to matrix lysis, as occurs in periodontitis. To test this hypothesis, we used contracted and attached lattices as gingival lamina propria equivalents. In such 3D models, supplementation of elastin peptides and plasminogen triggered an MMP-3/MMP-1 activation cascade and significant down-regulation of TIMPs production, further leading to intense collagen degradation. We propose that elastolysis, as occurs in periodontitis, potentiates collagenolysis, thus promoting disease progression.
Key Words: elastin • matrix metalloproteinase • stromelysin-1 • tissue inhibitor of metalloproteinase • collagen lattice
The elastic fiber system, composed of elastin and several microfibrillar constituents, has been identified in human gingiva (Mariotti, 1993), and the expression of tropoelastin by gingival fibroblasts has been established (Tsuruga et al., 2002). The number of polymorphonuclear neutrophils in the human gingival sulcus increases in patients with periodontal diseases (Van Dyke and Vaikuntam, 1994). During inflammatory phases, these cells release from their granules several elastolytic proteases, such as elastase, cathepsin G, and gelatinase B (Bieth, 1986). Within inflamed tissue, levels of active enzymes have been found to exceed those of their natural inhibitors, either serpins or tissue inhibitors of metalloproteinases (TIMPs) (Haerian et al., 1995; Meyer et al., 1997). These neutral proteinases, alone or in synergy, could degrade ex vivo the macromolecular constituents of the elastic fiber system and, thus, could generate locally the production of elastin-derived peptides (Berton et al., 2000). These peptides display chemotactic activity toward monocytes, pro-angiogenic activity, and modulate ion fluxes (Hornebeck et al., 2002; Robinet et al., 2005). All effects are mediated through the occupancy of an elastin-binding protein at the cell plasma membrane (Hinek et al., 1993), further identified as an alternatively spliced form of β-galactosidase (S-Gal) (Privitera et al., 1998). We previously demonstrated that binding of elastin peptides with the GXXPG consensus sequence to S-Gal induced matrix metalloproteinase (MMP) expression in several normal and cancer cells (Hornebeck et al., 2002), and thus we hypothesized that these peptides might additionally trigger MMP activation and an MMP escape mechanism from TIMP control, leading to matrix destruction as occurs in periodontitis (Birkedal-Hansen, 1993; Baker et al., 2002). To test this hypothesis, we evaluated their influence using human gingival fibroblasts in collagen lattices as a gingival connective tissue model.
Cell Cultures and Morphological Studies Gingival biopsies originated from individuals free of periodontal disease, during pre-prosthetic surgery. All subjects (aged 40 to 60 yrs old) gave their informed consent according to the Declaration of Helsinki, and the study was approved by the local ethical committee. Human gingival fibroblasts (4th–7th passages) were grown from explants in Dulbecco’s Eagle’s medium (DMEM, Gibco, Invitrogen, Eragny/Oise, France) supplemented with 10% fetal calf serum, and cultured at 37°C in humidified air containing 5% CO2. Fibroblasts were seeded at a concentration corresponding to cell confluence (200,000 cells) onto 24-well plates in 1 mL of DMEM lacking serum. Three-dimensional lattice cultures (floating lattices or attached lattices) were performed with the use of rat-tail acid-soluble type I collagen (Lorimier et al., 1996). Collagen-attached lattices were obtained with the use of nylon bars (Polylabo, Strasbourg, France) preventing the retraction of collagen lattices by cells (Lambert et al., 1992). Cells were counted with a Neubauer hemocytometer, and their DNA content was quantified by fluorometric assay (Lorimier et al., 1996). For any cell strain, triplicate experiments were performed. The lattices were prepared for scanning electron microscopy (SEM; JEOL JSM 5400LV) as previously described (Lorimier et al., 1996).
S-Gal Identification in Fibroblast Cultures We plated 100,000 gingival fibroblasts onto plastic dishes, in 1 mL DMEM containing 10% (v/v) serum, and incubated them for 24 hrs at 37°C. Cell plasma membranes were isolated as described (Brassart et al., 1998), and S-Gal was identified by Western blotting.
Fibroblast Activation by Elastin Peptides and Biochemical Assays For RT-PCR analyses, total RNA was isolated from 2 x 105 human gingival fibroblasts cultured for 3 hrs as monolayers in the presence or absence of elastin peptides, with Trizol reagent (Life Technologies, Paisley, UK). Total RNA was reverse-transcribed according to the manufacturer’s instructions (Life Technologies, no. 8025SA). The primers for MMPs, urokinase, TIMPs-1 and -2, and 18S RNA were synthesized by Invitrogen Life Technologies. The sequences of primers are specified in Appendix 2. cDNA products were amplified for 28, 30, and 32 cycles for all assays. For the identification of proteases and TIMPs, conditioned culture media were subjected to electrophoresis on 10% SDS-polyacrylamide substrate gels containing gelatin, or by reverse zymography, respectively (Brassart et al., 1998). For Western blotting, we separated proteins from conditioned media by 10% SDS-PAGE and blotted them onto nitrocellulose membrane (Immobilon P, Millipore, Bedford, MA, USA) (Brassart et al., 1998). The transferred proteins were probed with primary antibodies (1:1000 monoclonal anti-human MMP-1 [Ab-1], MMP-3 [Ab-1], TIMP-1 [Ab-2], and TIMP-2 [Ab-4]; Oncogen Research Products, Boston, MA, US) and secondary antibody (goat anti-mouse IgG [H+L]-peroxidase, 1:10,000; Immunotech, Marseille, France). Blots were developed by chemiluminescence assay (Kit ECL/RPN 2069, Amersham Pharmacia Biotech, Orsay, France).We determined the degradation of the collagen lattice by quantifying the 4-hydroxyproline content in conditioned culture media (Lorimier et al., 1996).
Statistical Analyses
Stimulation of MMP-3 Expression in Gingival Fibroblasts by Elastin Peptides is Receptor-mediated We first examined the influence of elastin peptides on protease (urokinase-plasminogen activator [uPA], MMPs) and TIMPs expression by RT-PCR analysis, using gingival fibroblasts grown on plastic dishes for 3 hrs. In the presence of peptides, levels of uPA, MMP-14, TIMP-1, and TIMP-2 mRNAs remained nearly constant; MMP-1 and MMP-2 expression was only slightly elevated, and only trace levels of MMP-13 mRNA were detected. In contrast, MMP-3 expression was significantly enhanced (Fig. 1A  ).
Elastin peptides stimulated MMP-3 production (3.5-fold average increase with 6 different cell strains; Fig. 1B  ). MMP-1 secretion was also increased, but to a lesser extent, and MMP-2 levels were not statistically different in the presence and absence of peptides. Whatever our experimental conditions, only MMP zymogen forms were detected (not shown).
S-Gal expression by gingival fibroblasts was demonstrated by Western blot, where only 1 protein band, with a 64-kDa apparent molecular weight, was identified (Fig. 1CI
Influence of Elastin Peptides on the Capacity of Fibroblasts to Degrade Floating Collagen Lattices
Elastin peptides in the presence of plasminogen exerted a major influence on pro-MMP-3 and pro-MMP-1 activation (Figs. 2B, 2C ). Under these conditions, pro-MMP-2 activation was affected only minimally (not shown). Strikingly, these peptides, in combination with plasminogen, decreased the levels of both TIMP-1 and TIMP-2 (Fig. 2D). A similar, but less pronounced, effect was observed with plasminogen (Plgn) alone. Also, an elastin-peptide- and plasminogen-mediated decrease in TIMP-1 and TIMP-2 production was confirmed for the proteins (Figs. 2E, 2F). At day 4, nearly one-quarter of collagen was degraded by the combined action of plasminogen and elastin peptides (Fig. 2G), which represented a three-fold enhancement in the rate of collagenolysis, as compared with lattices containing plasminogen only.
Elastin-mediated Collagenolysis in Gingival-fibroblast-populated Collagen-attached Lattices
Overexpression of matrix metalloproteinases is a hallmark of tissue destruction in oral diseases (Birkedal-Hansen, 1993; Sorsa et al., 2004). Among this family of proteinases, stromelysin-1 (MMP-3) displays an essential function in tissue homeostasis, since it cleaves a wide repertoire of matrix constituents and is considered an important activator of collagenase-1 (MMP-1) (Unemori et al., 1991). In a cross-sectional study of gingival crevicular fluid, MMP-3 and TIMP-1 levels were described as agents discriminating between healthy and diseased periodontal sites (Haerian et al., 1995), and periodontal pockets containing high MMP-3 levels present higher risk for periodontal disease progression (Alpagot et al., 2001). We demonstrated here that elastin fragments, as might be locally released following elastase action at inflammatory sites, up-regulated MMP-3 expression in human gingival fibroblast cultures. Maximal effect was observed at 50 to 100 µg/mL concentrations of peptide, a level exceeding their concentration in the blood circulation (Hornebeck et al., 2002). However, these peptides might be locally more active, since fibroblasts were shown to locate close to elastic fibers from the gingiva (Tsugura et al., 2002). Although tropoelastin was described as also binding to  vβ3 (Rodgers and Weiss, 2004), most biological effects of elastin peptides are mediated through S-Gal binding (Hinek et al., 1993; Hornebeck et al., 2002). It needs to be considered that insoluble elastin within fibers is covered by a mantle of microfibrils, so that S-Gal binding and further cell activation necessitate elastolysis. We here showed that human gingival fibroblasts expressed S-Gal, thus confirming the ubiquitous character of this receptor. Since lactose and VGVAPG peptides could inhibit and reproduce, respectively, the effects of elastin peptides on MMP-3 overexpression, we assumed that it was S-Gal-mediated. In melanoma as well as endothelial cells, elastin binding to S-Gal up-regulated MT1-MMP and MMP-2 expression and activation, while having no or minimal influence on the production of other members of the Matrixin family (Robinet et al., 2005). In dermal fibroblast cultures, elastin peptides were shown to induce, to similar extents, MMP-1 and MMP-3 expression (Duca et al., 2002), while they appeared to act more specifically on MMP-3 expression levels in their gingival counterparts. The reasons for such differences in elastin-peptide-mediated MMP expression among cell types are unresolved, and, in most instances, enzyme induction is associated with activation of the ERK1/2 signaling pathway (Duca et al., 2002). Triggering the plasmin/MMP-3/MMP-1 proteolytic cascade necessitates that (i) cells do express plasminogen activators, and (ii) the concentration of active enzymes exceeds the level of TIMPs produced by cells. Thus, we used gingival-fibroblast-populated type I collagen lattices, either free to retract or maintained under tension, as a model system to mimic the physiological three-dimensional environnement of the cell (Lambert et al., 1992; Ravanti et al., 1999). Such a gingival connective tissue model presents advantages, compared with its dermal counterpart, in keeping with the high constitutive expression of uPA by gingival fibroblasts, a key initiator enzyme in MMP proteolytic cascades (Lorimier et al., 1996). Therefore, we supplemented cell culture medium with plasminogen at a concentration close to that found in tissues (Vassalli et al., 1991), to generate plasmin, an important MMP activator. In floating collagen lattices, the elastin-peptide-enhancing influence on collagenolysis was clearly attributed to increased MMP-3-mediated MMP-1 activation, since levels of MMP-2 production and activation remained nearly constant in the presence or absence of peptides in plasminogen-containing cultures. Although MMP-13 was induced in this cell model system, elastin peptides had no influence on its expression (not shown). Such differences in enzyme expression among collagenases can be related to differences in signaling cascade activation: p38 activation is required for induction of MMP-13 (Ravanti et al., 1999), whereas S-Gal-mediated MMP-3-, as well as MMP-1-, enhanced expression is associated with the ERK1/2 pathway (Duca et al., 2002). Additionally, collagen degradation could be amplified by the TIMP-1 and TIMP-2 down-regulation in plasminogen and, more intensely, in plasminogen- and elastin-peptide-containing collagen gels. Analysis of those data argues for a TIMP post-transcriptional regulatory mechanism, since TIMPs mRNA steady-state levels were constant in the presence or absence of elastin peptides. We first suspected that degradation of TIMP-1 by generated plasmin was the underlying mechanism, but we confirmed that this inhibitor was not, or only weakly, hydrolyzed by plasmin (Okada et al., 1988). Culturing melanoma cells in a 3D-collagen environment also led to a conspicuous disappearance of TIMP-2 (Kurschat et al., 1999) that could be attributed to an MT1-MMP-dependent internalization and degradation of TIMP-2, as described in other tumor cell lines (Maquoi et al., 2000). Alternatively, TIMP-2 endocytosis by low-density-related lipoprotein receptors might be increased under these experimental conditions (Emonard et al., 2004). An attached collagen lattice was used as a 3D culture model for evaluating elastin peptides’ influence on collagenolysis. When cells are maintained under tension, they adopt an anabolic phenotype with repressed protease expression and collagen turnover (Lambert et al., 1992; Rosenfeldt and Grinnell, 2000). Indeed, as observed here, collagenolysis was low when lattices were grown in the absence of elastin peptides. In contrast, in the presence of peptides and plasminogen, cells behaved similarly as within contracted lattices, with the induction of MMP activation, down-regulation of TIMP-2 (not shown), and significant collagenolysis. Following 4 days of culture, these lattices detached from their supports, and retraction was observed, emphasizing the putative role of S-Gal as a mechano-receptor. Analysis of previous data demonstrated that levels of active leukocyte elastase in the crevicular fluid increased with periodontal disease progression (Meyer et al., 1997). Generated elastin peptides could act as potent catalysts of gingival connective tissue lysis—first, as amplifiers of elastase production by leukocytes, and second, as shown here, as inducers of an MMP-3/MMP-1 proteolytic cascade and collagenolysis in models of gingival connective tissue equivalents.
This work was supported by INSERM ERM0203 and funds provided by the French Ministery of Research (Grant 556021). The authors are endebted to Mrs. Martine Decarme and Christine Guillaume for helpful technical assistance.
A supplemental appendix to this article is published electronically only at http://www.dentalresearch.org. Received for publication July 29, 2005. Revision received April 24, 2006. Accepted for publication May 3, 3006.
Journal of Dental Research, Vol. 85, No. 8,
745-750 (2006) This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
