Binding of Pro-Matrix Metalloproteinase 9 by Fusobacterium nucleatum subsp. nucleatum as a Mechanism To Promote the Invasion of a Reconstituted Basement Membrane

The pathophysiology of periodontitis, a bacteria-induced inflammatory disease leading to tooth loss, is a complex process involving the participation of both bacterial and host proteolytic enzymes (9, 10). Studies have provided evidence that matrix metalloproteinases (MMPs) secreted by mammalian cells such as neutrophils, macrophages, and epithelial cells play a critical role in periodontal tissue destruction (10). These enzymes, which are produced as latent proenzymes, can cleave almost all the macromolecules of the extracellular matrix, including collagen, fibronectin, and proteoglycan. More specifically, MMP-9 (gelatinase B) from neutrophils was reported to be prominent in gingival crevicular fluid and gingival tissue specimens from patients with periodontitis (14).

Previous studies have shown that cell surface receptors of several bacterial pathogens can bind the 90-kDa plasma proenzyme plasminogen, which can be subsequently activated into plasmin, a broad-spectrum serine protease (7). In this study, we hypothesized that, like plasminogen, pro-MMPs could bind to the cell surface of bacteria, thus allowing them to acquire a proteolytic activity. To our knowledge, such a phenomenon has never been reported in the literature. Binding of pro-MMPs by periodontopathogens could increase tissue damage and promote their capacity to invade host tissues and disseminate to cause focal infections. The aim of our study was thus to investigate the binding and activation of human pro-MMP-9 on the cell surface of the periodontopathogenic bacterial species Fusobacterium nucleatum. The potential role of this acquired proteolytic activity in tissue invasion and destruction was evaluated using a reconstituted basement membrane (Matrigel).

Three subspecies of F. nucleatum (F. nucleatum subsp. nucleatum ATCC 25586 [isolated from human inflamed gingiva], F. nucleatum subsp. vincentii ATCC 49256 [isolated from human periodontal pocket], and F. nucleatum subsp. polymorphum ATCC 10953 [isolated from human cervico-facial lesion]) were grown in Todd-Hewitt broth (BBL Microbiology Systems, Cockeysville, Md.) at 37°C for 24 h under anaerobic conditions (N₂/H₂/CO₂, 80:10:10). Cells were harvested by centrifugation (10,000 × g for 10 min) and suspended in 50 mM phosphate-buffered saline (PBS), pH 7.2, to an optical density at 660 nm of 1.0. This corresponded to a concentration of 1 × 10⁹ bacteria/ml as determined by using a Petroff-Hausser counting chamber. The capacity of fusobacteria to bind human pro-MMP-9 was evaluated by an enzyme-linked immunosorbent assay (ELISA). One hundred microliters of the bacterial suspension was added to the wells of a microtiter plate (Nunc Maxisorp, Roskilde, Denmark), and bacteria were allowed to attach for 16 h at 37°C in a humid atmosphere. The liquid phase was then discarded and the wells were washed three times with PBS prior to fixing bacteria by adding 0.05% (vol/vol) glutaraldehyde for 30 min. The wells were again washed three times with PBS prior to blocking unreacted sites with 3% gelatin for 1 h. Human pro-MMP-9 (2 μg/100 μl; Oncogene, Cambridge, Mass.) was added to wells containing bacterial layers, and the plate was incubated for 2 h at room temperature. The wells were extensively washed three times with PBS prior to adding sheep anti-MMP-9 antibody (1:500; Biodesign, Kennebunk, Maine) as the first antibody (1 h) and then rabbit anti-sheep immunoglobulin G conjugated to alkaline phosphatase (1:5,000; Bethyl Laboratories, Montgomery, Tex.) as the second antibody (1 h). Three washings with 50 mM Tris hydrochloride buffer (pH 7.2) containing 0.05% Tween 20 were performed after the incubation with each antibody. A negative control in which wells were covered with 3% gelatin instead of bacteria was performed. An additional control was carried out by omitting the incubation step of bacteria with pro-MMP-9. The enzymatic reaction was revealed in 50 mM carbonate buffer (pH 9.8) containing p-nitrophenylphosphate, and the absorbance was measured at 405 nm after a 2-h incubation. Binding of p-aminophenylmercuric acetate (APMA; 0.5 mM)-activated MMP-9 was also tested. Binding assays performed in the presence of polymyxin B (1 mM) or the lysine analog ε-aminocaproic acid (10 mM) were also done. The means ± standard deviations of three independent experiments were calculated for all the binding assays. A standard curve in which increasing amounts of pro-MMP-9 (3.2, 16, 80, and 400 ng) were immobilized on the bottoms of the wells was constructed to estimate the amount of pro-MMP-9 bound to the fusobacteria. All three subspecies of F. nucleatum tested in the ELISA bound pro-MMP-9. From the standard curve, the relative amounts of pro-MMP-9 bound by F. nucleatum subsp. nucleatum, F. nucleatum subsp. vincentii, and F. nucleatum subsp. polymorphum were estimated at 144 ± 31, 103 ± 51, and 96 ± 28 ng, respectively. These quantities were bound by 2 × 10⁷ cells, since approximately 20% of the initial quantity of cells added into a well attached to the bottom of the well (data not shown). When the pro-MMP-9 was activated with APMA prior to the binding assay, comparable binding levels were obtained. No binding of pro-MMP-9 to gelatin-coated wells was observed. F. nucleatum subsp. nucleatum, which tends to bind pro-MMP-9 more efficiently than the two other subspecies, was chosen for subsequent analysis. The lysine analog ε-aminocaproic acid did not reduce the amount of pro-MMP-9 bound by the fusobacteria, thus suggesting that the plasminogen receptor previously reported (3) is not involved in binding of pro-MMP-9. Furthermore, the addition of polymyxin B, a lipid A-binding substance, did not reduce binding, suggesting that lipopolysaccharide is not involved.

The conversion of pro-MMP-9 bound to fusobacteria into a proteolytically active form was then investigated. The F. nucleatum subsp. nucleatum cell suspension was incubated with pro-MMP-9 (0.14 ng/μl) for 2 h at room temperature with gentle shaking. The cells were then washed three times in PBS by centrifugation at 10,000 × g for 10 min and suspended in the original volume of PBS. The pro-MMP-9 bound to the bacteria was activated by adding 0.5 mM APMA to the cell suspension. After a 30-min incubation, the cell-associated gelatinase activity was revealed by zymography on a 10% polyacrylamide gel containing 0.5% gelatin as previously described (4). Zymography revealed the presence of bands with gelatinase activity only with fusobacteria that had been incubated with pro-MMP-9 (Fig. 1A). Bands with a molecular mass corresponding to pro-MMP-9 (92 kDa) and to the predominant active forms of MMP-9 (82 and 65 kDa) were detected when pro-MMP-9-coated cells were treated with APMA. A colorimetric gelatinase assay (Boehringer Mannheim Canada, Laval, Quebec, Canada) was also used to confirm the acquisition of proteolytic activity by F. nucleatum subsp. nucleatum. In this assay, cells were incubated with biotin-labeled gelatin prior to measuring gelatinase activity in a streptavidin-coated microtiter plate according to the manufacturer's instructions. High gelatinase activity yielded a low signal as determined by the absorbance at 405 nm. Bacteria coated with pro-MMP-9 and subsequently treated with APMA had significant gelatinase activity compared to control cells (uncoated or coated with pro-MMP-9 without activation) (Fig. 1B).

To investigate the contribution of MMP-9 activity acquired by F. nucleatum subsp. nucleatum in invasion of a reconstituted basement membrane, ¹⁴C-labeled bacteria were prepared as described in a previous study (1). One hundred microliters of the bacterial suspension was added to EcoLite scintillation liquid (ICN, Costa Mesa, Calif.), and the radioactivity was counted using a scintillation counter. The optical density at 660 nm of the same radioactive cell suspension was measured, and a similar nonradioactive cell suspension was prepared to determine the bacterial concentration using a Petroff-Hausser counting chamber. Radiolabeled cells (uncoated, pro-MMP-9-coated, and APMA-activated MMP-9-coated) were then tested for their capacity to invade a reconstituted basement membrane (Matrigel; Sigma Chemical Co., St. Louis, Mo.). The Matrigel was diluted one-third in ice-cold PBS, and 100 μl of the dilution was placed on 8-μm-pore-size filters in 24-well Transwell cell culture chamber inserts (Costar, Cambridge, Mass.). The Matrigel was allowed to settle at 4°C for 30 min and was then gelled at 37°C for 24 h in an anaerobic chamber. The Matrigel was rehydrated by adding 100 μl of sterile, oxygen-free PBS for 1 h at 37°C in an anaerobic chamber. Excess PBS was removed, and 200 μl of ¹⁴C-labeled bacterial cells was added on the top of the gel. This corresponded to approximately 7 × 10⁸ bacteria (122,000 dpm) per well. In one assay, a 1:10 dilution of the suspension was used. Sterile, oxygen-free PBS (400 μl) was added to the lower wells of the chambers. The Transwell cell culture chambers were incubated at 37°C under anaerobiosis. To detect bacterial migration through the Matrigel, buffer in the lower well was removed after 2 and 4 h of incubation and placed in scintillation liquid, and the radioactivity was counted. The effect of treating MMP-9-coated fusobacteria at 60°C (30 min) or of adding 1 mM EDTA was investigated. The means ± standard deviations of three independent experiments were calculated. The radioactivity recovered in the lower well of the Transwell chamber showed that the fusobacteria had migrated through the reconstituted basement membrane. The results presented in Fig. 2 show that four times more ¹⁴C-labeled F. nucleatum subsp. nucleatum cells coated with APMA-activated MMP-9 penetrated the Matrigel after a 2-h incubation than did uncoated cells, and five times more penetrated after a 4-h incubation. Approximately 4.4% of the initial amount of MMP-9-coated cells deposited in the wells penetrated the Matrigel during the 4-h incubation period. Phase-contrast microscopy observation of the buffer recovered in the lower well of the Transwell chamber revealed the presence of intact fusobacteria for which the numbers were proportional to the radioactivity measured (data not shown). Heat treatment or the presence of EDTA prevented migration of bacteria coated with APMA-activated MMP-9. Lastly, the number of bacteria migrating through Matrigel was dependent on the amount applied on top of the gel.

In this study, we showed that F. nucleatum subsp. nucleatum, a gram-negative bacterium associated with gingivitis and chronic periodontitis (2), can bind human pro-MMP-9. To our knowledge, this is the first demonstration of the capacity of bacteria to attach a pro-MMP to their surface. The binding of MMP to the fusobacteria was strong, since it resisted extensive washing with PBS. Lipopolysaccharide does not appear to be involved in the binding, as shown by incorporating polymyxin B in the binding assay. The lysine analog ε-aminocaproic acid, which is known to inhibit the binding of plasminogen to F. nucleatum subsp. nucleatum (3), had no inhibitory effect. This suggests that different receptors on F. nucleatum subsp. nucleatum are responsible for the binding of plasminogen and pro-MMP-9. MMP-9 is known to possess a fibronectin-like sequence and a collagen-like domain (13). Interestingly, F. nucleatum cells attach to fibronectin and to a basement-like matrix (15). Also, pro-MMP-8, which does not possess a fibronectin-like sequence or collagen-like domain, was not bound by F. nucleatum subsp. nucleatum in the ELISA (data not shown).

MMPs, including MMP-9, are thought to play an important role in tissue destruction during periodontitis (10). Bacterial pathogens in the periodontal pocket may increase their virulence by taking advantage of their capacity to bind host-derived proteolytic enzymes such as MMPs. We have clearly demonstrated that pro-MMP-9 bound to F. nucleatum subsp. nucleatum could be activated to generate gelatinase activity on the cell surface of this bacterial species, which is normally a nonproteolytic bacterium. Furthermore, fusobacteria coated with active MMP-9 were more invasive than control bacteria in a basement membrane model. Given that periodontal pockets contain elevated levels of MMPs and their endogenous activators (10, 14), it is possible that in vivo, MMPs may be bound by F. nucleatum and be subsequently activated. The ability of F. nucleatum to acquire MMP activity could facilitate bacterial dissemination to surrounding periodontal tissues and blood vessels. Interestingly, MMP-9 has been reported to process interleukin-1β from its precursor form to an active form (12). The MMP-9 bound to F. nucleatum subsp. nucleatum may thus be involved in the regulation of the activity of cytokines, more particularly interleukin-1β, which plays a central role in acute and chronic inflammation.

Recent studies have suggested relationships between periodontitis and a number of systemic diseases and conditions, including cardiovascular diseases, respiratory tract infections, and premature births of low-weight babies (8). More particularly, there is an association between Fusobacterium infections in intact amniotic membranes and premature births (5, 11). F. nucleatum is isolated more frequently in amniotic fluid in women who give birth prematurely than any other bacteria (5). One of the hypotheses proposed to explain the presence of this bacterial species is that F. nucleatum migrates from other distant infection sites, such as the uterine cervix and the oral cavity. In fact, the species and subspecies isolated from amniotic fluid, including F. nucleatum subsp. vincentii and F. nucleatum subsp. nucleatum, are similar to those encountered in healthy and diseased subgingival sites, unlike the strains isolated from the inferior genital tract (5, 6). The capacity of F. nucleatum subsp. nucleatum to acquire MMP-9 activity may contribute to its capacity to migrate through tissues and cause focal infections.

• Open in new tab
• Download powerpoint

FIG. 1.

Gelatinase activity of F. nucleatum subsp. nucleatum ATCC 25586 following various treatments. (A) Analysis by zymography on gelatin-containing polyacrylamide gel. (B) Analysis by a colorimetric assay using biotin-labeled gelatin and for which a high gelatinase activity yields a low signal determined by measuring the absorbance at 405 nm. Lanes and columns: 1, APMA-activated pro-MMP-9 control; 2, bacteria coated with pro-MMP-9 subsequently activated with APMA; 3, bacteria coated with pro-MMP-9 without activation; 4, control uncoated bacteria. The asterisk indicates a significant difference between cells coated with APMA-activated pro-MMP-9 and control uncoated cells at a P value of < 0.005, determined by using a two-tailed Student's t test.

• Open in new tab
• Download powerpoint

FIG. 2.

Penetration of a reconstituted basement membrane (Matrigel) by F. nucleatum subsp. nucleatum ATCC 25586 under various conditions. Columns: 1, bacteria coated with APMA-activated pro-MMP-9; 2, bacteria (1/10 the amount applied in column 1) coated with APMA-activated pro-MMP-9; 3, control uncoated bacteria; 4, bacteria coated with APMA-activated pro-MMP-9 plus 1 mM EDTA; 5, bacteria coated with APMA-activated pro-MMP-9 and treated at 60°C for 30 min. After an incubation of 2 (left panel), and 4 h (right panel), radioactivity in the buffer present in the lower well was determined and used to estimate the number of cells that migrated through Matrigel. The values are the means ± standard deviations of three independent assays. The asterisk indicates differences between cells coated with APMA-activated pro-MMP-9 and control uncoated cells at P < 0.005, determined by using a two-tailed Student's t test.

• Copyright © 2004 American Society for Microbiology