INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Chronic periodontitis is recognized as one of the most common diseases afflicting humans. If left untreated it may result in tooth loss, and it is associated with an increase in the prevalence of ischemic heart diseases (29). Periodontitis is characterized by a chronic inflammatory process resulting in bone resorption, loss of tooth-supporting structures, and formation of periodontal pockets in response to the presence of bacteria. Porphyromonas gingivalis, a gram-negative, short-rod anaerobe, is consistently associated with chronic and severe disease in adults (16, 47). P. gingivalis synthesizes and secretes high levels of proteolytic enzymes involved in the pathogenicity of the bacterium. These proteinases have proved to be potent enzymes active against a wide range of substrates, such as matrix metalloproteinases, immunoglobulins, fibronectin, protease inhibitors, coagulation factors, and components of the complement and kallikrein-kinin cascade (1, 11, 15, 17, 20, 43). These bacterial proteases contribute to periodontal tissue destruction by a variety of mechanisms, including direct tissue degradation and modulation of host inflammatory responses (7, 51, 56). Among P. gingivalis proteinases, Arg-gingipain (gingipain-R) is a trypsin-like, cysteine proteinase found in soluble form in culture media or found associated with whole bacterial cells, of which at least two molecular species have been described and for which the crystal structure has been resolved (8, 14, 42, 43). Gingipain-R specifically cleaves polypeptides after arginyl residues and its enzymatic activity is not affected by plasma proteinase inhibitors, but normal saliva contains proteins, such as histatin 5, which may act as a substrate competitor and may control bacterial virulence (36, 39). P. gingivalis mutants made deficient in the gingipain-R genes are much less aggressive toward the host, and rodents or primates vaccinated against gingipain-R are protected against periodontitis, indicating that these enzymes are important virulence factors (35).

Chemokines are small secreted proteins that cause chemotactic migration of leukocytes by binding to specific receptors and play a major role in inflammatory processes (27). Based on a cysteine motif located near the N terminus of the protein, they have been subdivided into four families, of which the CXC and CC are the most numerous. Interleukin-8 (IL-8) belongs to the CXC family and specifically directs polymorphonuclear cell migration by binding to the CXCR1 and CXCR2 receptors (4). Gamma interferon (IFN-)-inducible protein 10 (IP-10) is also a CXC chemokine which lacks the ELR motif and binds to CXCR3, a receptor specifically expressed on activated T cells (24, 28). Upon binding to its receptor, IP-10 induces rapid, shear-resistant adhesion induction of effector cells (41). Thus, when released, the chemokines IL-8 and IP-10 recruit different cell types at sites of inflammation.

Periodontitis is characterized histologically by the presence of an abundant polymorphonuclear cell infiltrate in which mononuclear cells are also represented. T cells contribute to this inflammatory infiltrate and can be in close contact with resident fibroblasts (33). T cells are known to affect fibroblast metabolism by releasing soluble factors and, to a greater extent, by direct cell-to-cell contact (9, 32, 45, 48). In turn, human gingival fibroblasts (HGF) respond to bacterial and host stimuli by releasing prostaglandins, cytokines, and chemokines which may be directly or indirectly implicated in periodontal tissue destruction. Thus, the interplay between bacteria and host may lead to the development of chronic forms of periodontitis.

Gingipains have been shown to cleave and inactivate released cytokines and, as a consequence, are thought to impair the inflammatory response (7, 30, 55, 56). However, their activity on fibroblasts has received little attention. The aim of the present study was to elucidate the direct effect of P. gingivalis gingipain-R on chemokine production by HGF and its indirect effect on subsequent HGF responses to T-cell contact. We show that gingipain-R directly enhances IL-8 production while indirectly inhibiting IP-10 production. This may have an impact on the composition of inflammatory cell infiltrate in periodontitis since IL-8 preferentially recruits neutrophils and IP-10 attracts activated T cells.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents. Dulbecco's modified Eagle's medium (DMEM), nonessential amino acids, sodium pyruvate, and MEM-vitamins were purchased from Seromed (Biochrom KG, Berlin, Germany), and fetal calf serum (FCS), RPMI 1640 medium, penicillin, streptomycin, and TRIzol reagent were purchased from Gibco (Paisley, United Kingdom). Indomethacin, phorbol myristate acetate (PMA), NP-40, iodoacetamide, sucrose, phenylmethylsulfonyl fluoride, pepstatin, EDTA, polymyxin B sulfate, Na-benzoyl-DL-arginine b-naphthylamide (BApNA), and N-p-tosyl-L-lysine chloromethyl ketone (TLCK) were purchased from Sigma (St. Louis, Mo.), and dithiothreitol was purchased from Wako Pure Chemical Industries (Tokyo, Japan). Phytohemagglutinin (PHA) was from E-Y Laboratories Inc. (San Mateo, Calif.). [-³²P]dCTP (3,000 Ci/mmol) and [-³²P]UTP (3,000 Ci/mmol) were from Hartmann Analytic GmbH (Braunschweig, Germany). Leupeptin was from Fluka (Buchs, Switzerland). Human recombinant tumor necrosis factor alpha (TNF-) was from Biogen (Cambridge, Mass.). Soluble TNF receptor p55 was from Amgen (Thousand Oaks, Calif.), anti-IFN- was a gift from G. Garotta (Human Genome Sciences, Rockville, Md.), and IL-1Ra was from Synergen (Boulder, Colo.).

P. gingivalis culture and gingipain-R purification. P. gingivalis strain 381 (kindly provided by T. Nishihara and N. Hanada, Department of Oral Science, National Institute of Infectious Diseases, Tokyo, Japan) was cultured in Gifu anaerobic medium broth (Nissui, Tokyo, Japan) at 37°C for 24 h under anaerobic conditions (70% N₂, 15% H₂, 15% CO₂). P. gingivalis cysteine protease was purified according to the method described by Kontani et al. (22). Briefly, P. gingivalis cells were sonicated in phosphate-buffered 0.2% Triton X-100 and subjected to column chromatography on arginine Sepharose 4B, followed by DEAE Sepharose CL-6B, HiLoad 16/60 Sephacryl S-200HR, and HiTrap Q chromatography. The purified material was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blotting according to standard procedures. The enzyme activity was measured using the specific, synthetic BApNA substrate. The hydrolysis of the substrate was induced by the addition of 1 ml of BApNA (0.2 mM) in 50 mM Tris-HCl (pH 7.4) containing 0.2 mM dithiothreitol. The reaction was performed at 37°C for 10 min and monitored by the increase in absorbance at 410 nm. One unit of enzyme activity was defined as the amount of enzyme required to release 1 mM p-nitroanilide under these conditions. No lipopolysaccharide (LPS) contamination was detected in the purified gingipain-R by the Limulus amebocyte lysate. In some experiments, gingipain-R was used after heat inactivation at 80°C for 30 min. LPS was purified from the P. gingivalis 381 strain by the hot-phenol water method (52).

T-cell culture and plasma cell membrane preparation. Peripheral blood T lymphocytes (PBTL) were purified from buffy coats of healthy donors upon Ficoll-Paque gradient centrifugation and glass wool chromatography as previously described (49). PBTL from four distinct donors were used. PBTL were not pooled. They contained 94 to 98% CD2⁺, 83 to 94% CD3⁺, and <2% CD14⁺ as assessed by flow cytometry. The HUT-78 human cutaneous T lymphoma cell line was obtained from the American Type Culture Collection (Manassas, Va.). HUT-78 cells (10⁶ cells/ml) and PBTL (4 × 10⁶ cells/ml) were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated FCS, 50 mg of streptomycin/ml, 50 IU of penicillin/ml, and 2 mM L-glutamine for 18 h at 37°C in the absence or presence of 1 µg of PHA/ml and 5 ng of PMA/ml. After extensive washing in phosphate-buffered saline (PBS), T cells were suspended in PBS containing 0.68 M sucrose, 200 mM phenylmethylsulfonyl fluoride, 1 mM leupeptin, 0.1 mg of pepstatin/ml, and 5 mM EDTA and were sonicated (five 5-s bursts of 90 W each). The sonicated material was then centrifuged for 15 min at 4,000 × g to pellet nuclei and unbroken cells, and the supernatant was centrifuged for 45 min at 100,000 × g to obtain plasma cell membranes which were suspended at the theoretical concentration of 50 × 10⁶ cell equivalents/ml in PBS, 20 mM EDTA, and 5 mM iodoacetamide.

Human gingival fibroblast cultures. HGF were prepared from nine distinct biopsies of normal human gingival tissue obtained during surgical procedures after informed consent. The explants were cultured in 30-mm-diameter plastic dishes (Falcon) in DMEM supplemented with 10% FCS, 50 IU of penicillin/ml, 50 mg of streptomycin/ml, nonessential amino acids, and 1 mM sodium pyruvate at 37°C in 5% CO₂-air. Upon confluency, cells were trypsinized and transferred for further propagation. Fibroblasts were used from passages 3 to 10. HGF were seeded into 96-well flat-bottom plates (10⁴ cells/well) for microcultures and into 100-mm-diameter culture dishes (10⁶ cells/dish) for macrocultures. Cells were cultured in DMEM containing serum for 48 h and then in serum-free DMEM with or without 1.0 U of gingipain-R/ml for 12 h, unless otherwise stated. HGF were then washed with serum-free DMEM and further cultured in DMEM containing 1% FCS in the presence or absence of various stimuli for additional 12 h, unless otherwise stated. The culture supernatants were harvested and frozen at 80°C until analysis while the cells were washed twice with PBS and treated with 1% NP-40 for intracellular IL-8 determination. The cells in 100-mm-diameter culture dishes were treated with TRIzol reagent for analysis of mRNA expression.

Chemokine and PGE₂ determination. Commercial enzyme-linked immunosorbent assay (ELISA) kits were used to quantify IL-8 (CLB, Amsterdam, The Netherlands) and monocyte chemotactic protein-1 (MCP-1) and IP-10 (Hycult Biotechnology, Uden, The Netherlands). Cell-associated IL-8 was quantified upon suspension in 100 µl of 1% NP-40 in PBS. The reference standards were also diluted in 1% NP-40 in PBS. Prostaglandin-E₂ (PGE₂) was measured with a double-antibody radio-immunoassay as previously described (5).

RNA extraction and RPA analysis. Total RNA was isolated by lysing HGF with TRIzol reagent (Life Technologies) according to the manufacturer's instructions and analyzed for the levels of expression of IL-8, MCP-1, IP-10, RANTES, MIP-1, MIP-1, I-309, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNAs by RiboQuant RNase protection assay (RPA) using the hCK-5 multiprobe template set from PharMingen (San Diego, Calif.). Briefly, riboprobes were ³²P labeled and hybridized overnight in solution with 10 µg of the RNA samples. The hybridized RNA was digested with RNase, and the remaining "RNase-protected" probes were purified, resolved on denaturating polyacrylamide gels, and imaged by autoradiography according to the RiboQuant protocol. Autoradiography was quantified using the ImageQuant Software (Molecular Dynamics). The cyclooxygenase-2 (COX-2) mRNA level was determined by Northern blot analysis as previously described (9) using a specific cDNA probe kindly provided by P. E. Poubelle (University of Laval, Quebec, Canada).

Statistical analysis. Statistical evaluation was performed by Student's t test for paired data, and data were considered significant at a P value of <0.05.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Gingipain-R. The purified protease preparation used in this study was identified as a single band with a molecular mass of 47 kDa by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and was specifically recognized by an anti-gingipain-R antibody (kindly provided by M. Nishikata, Hokkaido University, Sapporo, Japan) on a Western blot (not shown). The protease showed trypsin-like activity using BApNA as substrate, with optimal activity at pH 7.0. The activity of the P. gingivalis protease was strongly inhibited by leupeptin and TLCK, since 0.1 mM leupeptin abrogated its activity and 0.1 mM TLCK reduced it by 75%. The amino-terminal sequence of the purified material was Y-T-P-V-E-E-K-E-N-G-R-M-I-V-I-V-A-K-K-Y, which corresponds to the published sequence of gingipain-R (14, 40).

Effect of gingipain-R on IL-8 production by HGF. Periodontitis is characterized by an accumulation of polymorphonuclear cells within the periodontal tissues and in the crevicular fluid. One of the most important chemoattractants of polymorphonuclear cells is IL-8. To ascertain whether gingipain-R could affect the production of IL-8, HGF were treated with gingipain-R and IL-8 mRNA was assessed by RPA. IL-8 mRNA steady-state levels were observed to increase in a time-dependent manner substantially more in gingipain-R-treated HGF than in untreated HGF (Fig. 1B). However, when IL-8 protein was measured in 12-h-culture fluids, no significant differences were observed between treated and untreated HGF, while intracellular IL-8 was increased in gingipain-R-treated HGF (Fig. 1C). Other authors have observed that gingipain-R may degrade IL-8 protein (30, 56). Thus, the discrepancy between mRNA and extracellular protein IL-8 levels in cultures treated with gingipain-R could be due to enzymatic degradation of IL-8 protein. Indeed, when we treated recombinant IL-8 protein with gingipain-R, we observed a time-dependent loss in IL-8 immunoreactivity detected by ELISA (not shown). To circumvent the degradation of IL-8 protein by gingipain-R, HGF were treated with gingipain-R for 12 h, and then the culture medium was replaced and IL-8 was measured after 12 additional hours of culture (Fig. 2A). Under this condition, gingipain-R-treated HGF yielded levels of IL-8 that were significantly higher than those of untreated HGF (Fig. 2B and C). This indicates that gingipain-R enhances both mRNA and protein levels of IL-8 in HGF, although IL-8 protein immunoreactivity is decreased by gingipain-R.

View larger version (20K): [in this window] [in a new window]   FIG. 1.   Effect of gingipain-R on IL-8 mRNA steady-state and IL-8 protein levels in gingival fibroblasts. (A) Experiment flow chart. HGF (104/well) were plated 48 h before gingipain-R (1 U/ml) was or was not added, and HGF were harvested at the indicated time for RPA (B) or ELISA (C). (B) IL-8 mRNA was detected by RPA. Results are from one of two distinct experiments with similar results. (C) IL-8 protein. Levels of cell-associated IL-8 and IL-8 released in culture supernatants were assessed by ELISA. Columns and error bars represent the mean and standard deviation, respectively, of four independent experiments. ns, nonstimulated.

View larger version (32K): [in this window] [in a new window]   FIG. 2.   Effect of T-cell contact on IL-8 production by HGF previously treated with gingipain-R. (A) Experiment flow chart. (B and C) HGF (passages 3 to 8) were treated with gingipain-R (1 U/ml) or cultured in medium alone (untreated) for 12 h and then were activated for 12 h by plasma cell membranes from nonstimulated (ns) or stimulated (s) HUT-78 T cells at 5 × 104 cell equivalents per well (B) or nonstimulated or stimulated peripheral blood T lymphocytes (PBTL) at 105 cell equivalents per well (C). (D) HGF were treated with gingipain-R (1 U/ml) for various times as indicated and stimulated by HUT-78s cells (5 × 104 cells/well) for 12 h. (E) HGF were treated for 12 h with the indicated doses of gingipain-R and then activated by HUT-78s (5 × 104 cell/well) for 12 h. IL-8 was assessed in supernatants by ELISA. For all shown data, T cells were stimulated by PHA plus PMA as described in Materials and Methods. Columns represent the mean and standard deviation of eight (B) and four (C) distinct experiments, respectively. Figures on top of columns indicate the P value when comparing the responses of gingipain-R-treated HGF to those of untreated HGF.

Effect of gingipain-R on HGF and their subsequent response to T-cell contact. We next determined whether gingipain-R-treated HGF still had the capacity to respond to T-cell contact. Indeed, T cells are present in periodontitis and are known to be powerful activators of fibroblasts (6, 32). In order to minimize the effect of released soluble mediators and to mimic T-cell contact, we used purified plasma cell membranes instead of intact cells as T-cell effectors. An increase in IL-8 levels was observed when HGF were activated by T-cell contact (Fig. 2B and C). Of interest, IL-8 production was consistently and significantly higher when HGF were treated with gingipain-R before T-cell contact. The enhanced IL-8 production by gingipain-R-treated HGF was observed with cell membranes from HUT-78 cells and from peripheral blood T cells. IL-8 levels were higher when cell membranes were derived from activated compared to nonactivated T cells (Fig. 2B and C). Of further interest, both time of treatment and dose of gingipain-R affected the subsequent response of HGF to T-cell contact. Indeed, IL-8 production by HGF exposed to gingipain-R steadily enhanced from 1 to 6 h and leveled out after 6 h (Fig. 2D). A dose of gingipain-R as low as 0.01 U/ml sufficed to enhance IL-8 production over the baseline, and this effect increased in a dose-dependent manner up to 1 U/ml (Fig. 2E). Higher doses of gingipain-R resulted in HGF detachment from culture dishes. All together, these data indicate that gingipain-R robustly primes HGF for higher production of IL-8.

Effect of endogenous PGE₂ on IL-8 production by HGF. We next investigated whether gingipain-R would modulate the production of PGE₂, since PGE₂ was shown to enhance IL-8 production by fibroblasts (2). Spontaneous and T-cell-induced PGE₂ levels were higher in HGF exposed to gingipain-R than in unexposed HGF, and indomethacin abrogated PGE₂ production (Fig. 3A). Concomitant gingipain-R-dependent IL-8 production was significantly reduced when indomethacin was added to the cultures while indomethacin did not affect IL-8 production in HGF not treated with gingipain-R (Fig. 3B). Indomethacin is a nonselective inhibitor of both inducible (COX-2) and noninducible (COX-1) cyclooxygenases which catalyze the synthesis of PGE₂. We therefore analyzed whether gingipain-R would affect the expression levels of COX-2 mRNA. COX-2 mRNA was observed in HGF when they were exposed to gingipain-R (Fig. 3C). COX-2 steady-state mRNA levels were strongly enhanced after 2 h and were still detectable at 6 h (Fig. 3C) and 12 h (Fig. 3D) following treatment with gingipain-R. When untreated HGF were activated by T-cell contact, COX-2 mRNA levels became marked within 30 min. No significant additional increase in COX-2 mRNA levels was observed in gingipain-R-treated HGF upon T-cell contact when compared to untreated HGF (Fig. 3D and E). Overall, these results indicate that the enhancement of IL-8 production by gingipain-R treatment is, at least in part, due to PGE₂ production.

View larger version (41K): [in this window] [in a new window]   FIG. 3.   Effect of gingipain-R HGF treatment on PGE2 production and COX-2 steady-state mRNA levels. (A and B) Culture conditions were as described in the legend of Fig. 2, except that 105 M indomethacin was added in some culture conditions. Culture supernatants were analyzed for PGE2 (A) and IL-8 (B) content. Columns and error bars represent the mean and standard deviation of three independent experiments. (C) Confluent monolayers of HGF were cultured in medium alone or in the presence of 1 U of gingipain-R/ml for the indicated time. Total RNA was then extracted and Northern blot analysis was performed using COX-2 and GAPDH probes, both exposed for 4 h. (D) Confluent monolayers of human gingival fibroblasts were cultured with or without gingipain-R for 12 h and then cultured for another 0.5 h with stimulated HUT-78 cell membranes (2.5 × 106 cell equivalents per condition). Northern blot analysis was performed as described for panel C. (E) COX-2 mRNA quantification by Image Quant software was assessed on HGF upon activation by HUT-78 cell membranes. (C, D, and E) Results shown are from one of two distinct experiments with similar results.

Gingipain-R proteolytic activity requirements for IL-8 enhancement. Proteases may affect cell responses through interaction with receptors which are activated upon partial proteolytic cleavage (proteinase-activated receptor [PAR]) (12). We therefore wished to investigate whether gingipain-R proteolytic activity was needed to enhance IL-8 production. The synthetic protease inhibitors leupeptin and TLCK as well as heat inactivation abrogated the capacity of gingipain-R to enhance both spontaneous and T-cell-dependent IL-8 production by HGF (Fig. 4A). Furthermore, the substrate competitor histatin 5, which is normally present in human saliva, was also effective in abrogating in a dose-dependent manner the gingipain-R enhancing activity (Fig. 4B). Although the Limulus assay did not reveal LPS in purified gingipain-R, we tested whether residual LPS could participate, at least in part, to gingipain-R-dependent IL-8. As shown in Fig. 4C, polymyxin B did not affect the enhanced IL-8 production by HGF exposed to gingipain-R, but in parallel experiments, it was fully capable of inhibiting IL-8 production induced by P. gingivalis LPS (Fig. 4D). These data indicate that the proteolytic activity of gingipain-R is required for IL-8 enhancement and that IL-8 enhancement is not due to LPS contamination of gingipain-R preparation.

View larger version (32K): [in this window] [in a new window]   FIG. 4.   Effect of protease inhibitors and polymyxin B on IL-8 production. Culture conditions were the same as described for Fig. 2. The culture supernatants were collected after 12 h of culture and analyzed for IL-8 content by ELISA. Columns and error bars represent the mean and standard deviation of two distinct experiments. (A) Native as well as heat-inactivated gingipain-R (equivalent to 1 U of native gingipain-R/ml), 0.1 mM TLCK, and 0.1 mM leupeptin. (B) Histatin 5 was purified as previously described (36). (C) PBTLs were used as stimulus. (D) LPS and polymyxin B were added at 10 and 100 µg/ml, respectively. No gingipain-R was used.

Selective effects of gingipain-R on chemokine mRNA levels in HGF. We then investigated the capacity of gingipain-R to modulate the chemokine production by HGF in response to T-cell contact. To this aim, the mRNA levels of several chemokines, including IL-8, MCP-1, IP-10, RANTES, MIP-1, MIP-1, and I-309, were simultaneously examined by RPA. Of all the chemokines tested, only IL-8, MCP-1, and IP-10 mRNA was constantly detected in 10 distinct experiments performed. The IL-8 mRNA level was upregulated in untreated HGF within 0.5 h upon T-cell contact (Fig. 5A). In gingipain-R-treated HGF, IL-8 mRNA was already highly expressed at baseline and slightly increased upon T-cell contact (Fig. 5A). MCP-1 mRNA was abundantly expressed in HGF at baseline (not shown) and at time zero and was not significantly modulated either by gingipain-R treatment or by T-cell contact (Fig. 5A). In contrast, no IP-10 mRNA was detectable at baseline and at time zero in gingipain-R-treated or untreated HGF. However, IP-10 mRNA was detectable in T-cell contact-activated HGF, the signal becoming strongly evident at 4 h (Fig. 5B). Interestingly, previous treatment by gingipain-R inhibited steady-state IP-10 mRNA levels when HGF were activated by T cells (Fig. 5B). Since the IP-10 mRNA signal peaked relatively late, IP-10 mRNA was studied up to 12 h after HGF activation by T cells. As shown in Fig. 5B, the inhibitory effect of gingipain-R on IP-10 mRNA persisted for at least 12 h after HGF were activated by T cells. When MCP-1, IL-8, and IP-10 protein levels in the supernatants of four independent experiments were determined, IP-10 levels were significantly reduced while IL-8 was enhanced and MCP-1 was unaffected in gingipain-R-treated HGF compared to untreated HGF (Table 1). Overall, these results indicate that gingipain-R has the capacity to selectively and finely modulate chemokine production by HGF in response to T-cell contact.

View larger version (44K): [in this window] [in a new window]   FIG. 5.   Effect of gingipain-R treatment on steady-state mRNA levels of IL-8, MCP-1, and IP-10 in HGF upon T-cell stimulation. Confluent monolayers of gingival fibroblasts were cultured for 12 h with or without gingipain-R and exposed to stimulated HUT-78 membranes (2.5 × 106 cell equivalents per condition) for the indicated times. mRNA was detected by RPA. Gels were exposed for 4 h (A) or 2 h (B). Results are from one of two distinct experiments with similar results.

Involvement of T-cell membrane-associated TNF in IL-8 induction. The final aim was to determine which of the molecules present on T-cell membranes were responsible for IL-8-enhanced production by HGF. Since in our past work membrane-associated TNF proved to mediate many activating properties of T-cell contact on fibroblasts (6), we tested the effect of a TNF blocking agent (soluble TNF-R p55). TNF blockade abrogated IL-8 induction by T-cell membranes in both untreated and gingipain-R-treated HGF (Fig. 6). In additional experiments, the blockade of membrane-associated IFN- and IL-1 resulted in a slight decrease of IL-8 production (not shown). Of interest, soluble TNF was a potent inducer of IL-8 production by HGF, even more so on gingipain-R-treated HGF. IL-8 release induced by soluble TNF was completely abolished by the TNF blocking agent. Thus, both T-cell membrane-associated and soluble TNF appear to play a major role in IL-8 production by HGF.

View larger version (28K): [in this window] [in a new window]   FIG. 6.   Involvement of T-cell membrane-associated TNF in enhanced IL-8 production by HGF. Culture conditions were as described for Fig. 2 except that HUT-78 (5 × 104 cell equivalents/culture), PBTL (105 cell equivalents/culture), and 5 ng of soluble TNF/ml were treated with 108 M TNF-R p75 for 30 min at 4°C before adding the preparations to HGF monolayers. Columns and error bars represent the mean and standard deviation of four (medium and HUTs) or two (other conditions) distinct experiments. s, stimulated; ns, nonstimulated.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, we have investigated for the first time the relative contribution and reciprocal influence of gingipain-R, a virulence factor produced by P. gingivalis, and of T-cell contact on chemokine production by HGF. We based the experimental model on the sequence of events likely to occur early in periodontitis. We reasoned that, first, bacteria invade the epithelium and then bacterial products may diffuse, affecting gingival tissue and stromal cells (HGF). Then, inflammatory cells are recruited. Thus, when activated T cells reach the site of bacterial colonization, they come into contact with stromal cells already exposed to bacterial products. The data show that gingipain-R enhances IL-8 mRNA levels in HGF. This results in enhanced early accumulation of cell-associated IL-8 and, subsequently, higher IL-8 levels in culture supernatants than in controls. Besides, HGF quickly upregulate IL-8 mRNA and release IL-8 when activated by contact with T cells. Interestingly, when HGF are exposed to gingipain-R and then activated by T-cell contact, the levels of IL-8 detected in culture supernatants are higher than those in controls, which indicates an additive proinflammatory effect of bacterial product and host effector response. Based on the capacity of gingipain-R to inactivate IL-8 and other cytokines, some authors have speculated that this bacterial enzyme decreases host inflammatory responses, thus contributing to chronic disease (7, 55, 56). Further, virulent P. gingivalis strains have been shown to block IL-8 production by gingival epithelial cells (10). However, Mikolajczyk and coauthors have demonstrated that soluble gingipain-R may sequentially enhance and then decrease the biological activity of IL-8 (30). Our data are in partial agreement with these findings, since no enhanced accumulation of immunoreactive IL-8 was observed in HGF supernatants when gingipain-R was present in the culture fluids. However, upon treatment with and subsequent removal of gingipain-R, HGF released larger amounts of IL-8 than untreated HGF, indicating that, at least in terms of IL-8 production, gingipain-R provides a potentiating rather than an attenuating stimulus, particularly when HGF are then activated by T-cell contact. We also show that the HGF capacity to produce chemokines in response to T-cell contact is not exclusively potentiated by previous exposure to gingipain-R. Indeed, contrary to the synthesis of IL-8, that of IP-10 induced by T-cell contact was significantly reduced in gingipain-R-treated HGF. Both IL-8 and IP-10 are CXC chemokines, but they attract different cell types. IL-8 preferentially recruits neutrophils, while IP-10 preferentially recruits activated T cells (27). Thus, gingipain-R may affect not only the amount but also the quality of the inflammatory cells recruited in periodontitis since it may enhance neutrophil numbers by increasing IL-8 and may reduce the numbers of activated T cells by decreasing IP-10 production. Similarly, these data imply that intracellular signals initiated in HGF by T-cell contact are modulated in an exquisitely selective manner by previous exposure to gingipain-R.

Gingipain-R proteolytic activity on soluble proteins has been extensively investigated. However, its direct effects on fibroblasts and other cells have received little attention. Although modulation of adhesion molecules on HGF and calcium fluxes on neutrophils has been described (25, 53), this is, to the best of our knowledge, the first report documenting a direct effect of gingipain-R in terms of the capacity to induce and regulate chemokine production by HGF. Neutrophil calcium fluxes have been linked with the capacity of gingipain-R to cleave a peptide corresponding to the activation sequence of PAR-2 (25). PAR-2 is a member of an expanding family of receptors with seven transmembrane domains coupled to G proteins, of which four have been identified by molecular cloning (19, 21, 31, 37, 50). They are activated when proteolytic enzymes cleave at specific sites the N-terminal portion of the first extracellular domain of the receptor, thus making available a tethered ligand domain for binding to the cleaved receptor (reviewed in reference 12). Indeed, gingipain-R has a trypsin-like activity, and PAR-2 has been identified because it is activated by trypsin and tryptase but not by thrombin (31, 38). Tryptase has recently been proved mitogenic on murine and human lung fibroblasts but not on dermal fibroblasts (3, 46). On the other hand, thrombin has been reported to induce IL-6 and NF-IL-6 on HGF in which PAR-2 mRNA has not been detected (18). In addition, thrombin was shown to upregulate IL-8 production by lung fibroblasts via PAR-1 (26). Although we have not determined whether PARs mediate the gingipain-R effect on HGF, we have shown that the enhanced IL-8 production by HGF is dependent on the enzymatic activity of gingipain-R (Fig. 4). Indeed, heat inactivation, the trypsin-specific inhibitor TLCK, and the serine and thiol protease inhibitor leupeptin abrogated the enhanced IL-8 production induced by gingipain-R on HGF. Thus, the data obtained are consistent with the hypothesis that gingipain-R affects HGF by interacting with protease-activated receptor(s). Of interest, histatin 5, normally present in parotid and submandibular saliva and a substrate competitor of high affinity for gingipain-R, inhibited in a dose-dependent manner the enhanced gingipain-R-dependent IL-8 production by HGF (8, 36). Thus, normal saliva contains a protein that inhibits the activity of bacterial products capable of modulating the host inflammatory response.

Our data show that gingipain-R induces COX-2 mRNA and that PGE₂ levels are enhanced in culture supernatants of gingipain-R-treated HGF (Fig. 3). In addition, the presence of indomethacin inhibits the enhanced response to subsequent HGF activation by T-cell contact, indicating that one of the intracellular pathways triggered by gingipain-R leads to prostanoid production, and this contributes to the inflammatory activity induced by this bacterial enzyme. Whether alone or associated with TNF-, IL-1 has been shown to induce COX-2-dependent PGE₂ production in lung, gingival, and synovial fibroblasts (13, 44, 54). Our data indicate that not only gingipain-R but also T-cell contact induces COX-2 gene expression on HGF (Fig. 3). Since membrane-associated TNF carried most of the IL-8-inducing capacity of T-cell membranes (Fig. 6), it is reasonable to assume that membrane-associated TNF is the main mediator involved in COX-2 mRNA upregulation when HGF are contact activated by T cells. This adds to the effects elicited on fibroblasts by T-cell membrane-associated TNF, which has been shown to be the main inducer of MMP-1 production upon T-cell contact (6). In our in vitro system, plasma cell membranes rather than intact T cells were used as effectors in mediating cell contact. Although outside the scope of the present study, previous experiments performed in our laboratory have documented that fibroblast responses induced by T-cell membranes are mimicked by activated, paraformaldehyde-fixed T cells and that plasma cell membrane preparations as opposed to nuclear or cytosolic preparations carry the biological activity (9, 23, 45). The induction of inflammatory cytokine mRNA on HGF upon interaction with living T cells has also been observed by others (32-34).

In conclusion, the data presented reveal new and unsuspected interactions, which are summarized in Table 2, between bacterial products with enzymatic activity and host effector functions. These interactions have a major impact on inflammatory responses and may be relevant to the mechanisms leading to periodontitis or favoring its chronic course.