Insertional Inactivation of pac and rmlB Genes Reduces the Release of Tumor Necrosis Factor Alpha, Interleukin-6, and Interleukin-8 Induced by Streptococcus mutans in Monocytic, Dental Pulp, and Periodontal Ligament Cells

Abstract

Streptococcus mutans possesses different cell wall molecules, such as protein of the I/II family, the serotype f polysaccharide rhamnose glucose polymer (RGP), and lipoteichoic acid (LTA), which act as adhesins and modulins, allowing S. mutans to colonize teeth and cause dental caries and pulpitis. We tested several isogenic mutants of S. mutans defective in protein I/II and/or RGP, as well as purified modulins such as protein I/II, RGP, and LTA, for their binding and activation abilities on monocytic, dental pulp (DP), and periodontal ligament (PDL) cells. Our results demonstrate that both protein I/II and RGP play important roles in streptococcal adherence to human monocytic and fibroblastic cells, whereas LTA is only a minor adhesin. In the activation process, the cytokine response elicited is polarized toward a Th1 response which seems principally due to protein I/II and RGP. Even if protein I/II seems to be more efficient in its purified form in triggering cells to release interleukin-8 (IL-8), RGP is the most efficient cytokine-stimulating component in intact bacteria, while LTA plays only a minor role. In cell activation, we showed, by using either cytochalasin D or coated ligands, that internalization of either S. mutans, S. mutans isogenic mutants, or purified ligands is not necessary to trigger cells to release IL-8. We also showed that, besides the implication of monocytes in pulpal inflammation, fibroblast-like cells such as DP and PDL cells are also actively implicated in local inflammation and in the generation of a Th1 response after stimulation with S. mutans cells or antigens.

Streptococcus mutans has been strongly associated with the initiation and propagation of human dental coronal surface caries (39). Furthermore, S. mutans invasion of dental pulp (DP), which occurs in several ways, including penetration through opened dentinal tubules (1), may contribute to the development of acute or chronic pulpitis and root surface caries (7, 15). Evidence suggests that as caries advances, S. mutans comes into contact with DP cells, including odontoblasts, endothelial cells, fibroblasts, monocytes/macrophages, and lymphocytes, and subsequently with periodontal ligament (PDL) cells, inducing inflammation and resulting injury (50). S. mutans possesses numerous cell surface molecules (20) which act as virulence factors that allow it to (i) adhere to salivary pellicles, (ii) accumulate within the dental biofilm, (iii) produce acids, (iv) invade dentinal tubules, (v) enter and invade DP, (vi) interact with pulpal cells, and finally (vii) interact with PDL cells after periapical diffusion. Exposure of host cells to S. mutans modulates cell activities such as up-regulation of cytokine synthesis or phenotype change (4, 18). Initial studies have shown that among the various S. mutans molecules, proteins of the I/II family, the serotype f polysaccharide rhamnose glucose polymer (RGP), and lipoteichoïc acid (LTA) function either as adhesins enabling S. mutans to participate in tooth colonization and invasion (19, 27, 33) or as modulins triggering cell functions (34, 35, 38, 45, 48) after binding to their cognate receptors.

Multiple evidence has shown that protein I/II, a cell wall-anchored, multiligand binding adhesin with a high molecular weight, is implicated in the adhesion of S. mutans to numerous salivary glycoproteins either in the fluid phase or when they are adsorbed onto hydroxyapatite. This protein has been implicated in (i) the initial step of bacterial colonization and (ii) interspecies coaggregation or agglutination (10, 19, 30). Protein I/II is also essential for invasion or coinvasion of dentinal tubules through its interaction with collagen type I (23, 24). Klein and colleagues have demonstrated the ability of protein I/II to operate as a modulin promoting the induction of proinflammatory cytokine synthesis after binding in a lectin-like mode of recognition to its cognate receptor on epithelial and endothelial cells, monocytes, and synoviocytes (17, 34, 35, 45) as well as the up-regulation of E-selectin, ICAM-1, and VCAM-1 expression on endothelial cells (46). It has also been shown recently that recognition of α₅β₁ integrin on endothelial cells is responsible for the production of interleukin-8 (IL-8) (5).

The serotype f polysaccharide RGP acts as a putative adhesin for the binding of S. mutans to tooth surfaces (27), heart muscle, and kidney tissues (37). RGP triggers various cells such as monocytes, endothelial cells, and epithelial cells, promoting proinflammatory cytokine release (35, 45), up-regulation of RFc γ (8), and production of NO in rat aortic cells (25). Furthermore, RGP binds to CD14 and CR3, but only the binding to CD14 has been correlated with the release of cytokines (35). Recently, Tsuda et al. (43) showed that the hydrophilic nature of RGP plays an important role in the resistance of S. mutans to phagocytosis by human polymorphonuclear leukocytes.

The microamphiphile LTA, anchored in the cytoplasmic membranes of gram-positive bacteria, exhibits many biological activities and can trigger various cells to induce the production of proinflammatory cytokines and NO (14, 38) in a CD14-dependent manner. Recently, Sugawara et al. (38) showed that purified LTA from Streptococcus sanguis or Bacillus subtilis acts, respectively, as an antagonist or agonist of lipopolysaccharide on human gingival fibroblasts. Furthermore, it has been postulated that LTA might play an important role in pulpitis by inducing apoptosis of human DP, which is suppressed by caspase inhibitors (48).

Although the importance of protein I/II in adherence to the tooth surface is well documented, little is known about the importance of protein I/II, RGP, and LTA in pulpal and periapical injury. The present paper describes a functional analysis of S. mutans protein I/II, RGP, and LTA with regard to their abilities to stimulate key pulpal and periodontal cells to release IL-6, IL-8, and IL-10. Isogenic mutants, constructed by insertional inactivation mutations, which are defective either in protein I/II, in RGP, or in protein I/II-RGP expression were used to determine the relative importance of each component in either cell binding or induction. The results show that (i) protein I/II and RGP are major adhesins responsible for the binding of S. mutans to human monocytes and dental cells, (ii) LTA acts only as a minor S. mutans adhesin, (iii) the binding of protein I/II and RGP to their cognate receptors activates human cells to release cytokine, (iv) RGP seems to be more accessible on the cell surfaces of whole bacteria in promoting cell activation, and (v) S. mutans LTA is a less potent inducer.

MATERIALS AND METHODS

Reagents.

Polymyxin B, cytochalasin D, pronase, vitamin D₃, lipopolysaccharide (LPS) from Escherichia coli O55:B5, and LTA from S. mutans were from Sigma-Aldrich (St. Quentin Fallavier, France). The enzyme immunoassay kit for tumor necrosis factor alpha (TNF-α), IL-6, IL-8, and IL-10 was from Biosource (Fleurus, Belgium), and N-hydroxysulfosuccinimide-long-chain biotin was from Pierce (Rockford, Ill.). Rabbit anti-Streptococcus pyogenes LTA immunoglobulin G (IgG) was from Biogenesis (Poole, England). Alkaline phosphatase (AP)-labeled anti-rabbit IgG and mouse monoclonal anti-CD14 (TÜK4) antibodies were from Dako S.A. (Trappes, France), and AP-streptavidin, RPMI 1640, fetal calf serum, l-glutamine, penicillin, and streptomycin were from Gibco BRL (Life Technologies, Cergy Pontoise, France). Cell culture media and LTA had endotoxin contents that never exceeded 0.004 ng per ml, as tested by a Limulus chromogenic assay. Rabbit anti-protein I/IIf from S. mutans KT6219 IgG was prepared as previously described (33). Throughout this study, buffers were prepared with apyrogenic water.

Bacterial strains, DNA manipulation, and characteristics of the isogenic mutants.

S. mutans KT6219 and its RGP-defective mutant, KT6219DR, were selected from the stock collection of the Department of Preventive Dentistry, Kyushu University, Faculty of Dental Science, Fukuoka, Japan. The pac gene coding for protein I/II in KT6219 was interrupted by the spectinomycin resistance gene from Enterococcus faecalis. Briefly, a 1.7-kb SpeI-EcoRI fragment of the pac gene of strain MT8148 was cloned into pUC119, and a 1.9-kb fragment containing both the spectinomycin resistance gene and the P15A replication origin was inserted into the BsmI site of the 1.7-kb fragment. The resultant fragment was detached from pUC119 and introduced into KT6219 as previously described (43). Transformants were selected on a tryptic soy agar plate containing 250 μg of spectinomycin per ml. The defect in protein I/II production was confirmed by Western blot analysis, and the transformant was designated KT6219-11. To construct a double mutant defective in both protein I/II and RGP production, chromosomal DNA of KT6219DR was introduced into KT6219-11 by the same procedure as that used for KT6219-11 construction, and transformants were selected on a tryptic soy agar plate containing 10 μg of erythromycin per ml. The loss of RGP production in cells was checked by high-performance liquid chromatographic analysis of cell wall polysaccharides as previously described (44), and the transformant was designated KT6219DR-11.

Purification of protein I/IIf and RGP.

Bacterial cells were grown anaerobically in brain heart infusion broth (Difco Laboratories, Detroit, Mich.). In some instances bacterial cells were heated at 60°C for 30 min or incubated at 37°C with 50 mM NaN₃ for 3 h. Wall-extracted antigens were prepared as described by Soell et al. (34), and protein I/IIf from S. mutans KT6219 was further purified from wall-extracted antigens by gel filtration and immunoaffinity chromatography as previously described (2). The purity of protein I/IIf from KT6219 was checked by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) after staining with Coomassie blue. Protein I/IIf migrated as a single band with an apparent molecular mass of 195 kDa which reacts with anti-I/IIf antibodies (Fig. 1).

FIG. 1.

(Left) SDS-PAGE of purified protein I/IIf (lane 1) and RGP (lane 2) stained with Coomassie blue. (Right) Immunoblot analysis of purified protein I/IIf (lane 1) and RGP (lane 2) with polyclonal antibodies raised against S. mutans protein I/IIf. Molecular weight markers are indicated on the left.

The serotype polysaccharide antigen of S. mutans KT6219, formed from RGP, was prepared by the method described by Benabdelmoumene et al. (8), and SDS-PAGE analysis showed the absence of Coomassie blue staining materials (Fig. 1).

Preparation of cells and solubilization of membrane proteins.

THP-1 cells were cultivated as previously described (11). Human PDL cells were obtained from erupted premolars from healthy subjects (4), grown in RPMI 1640 containing 2 mM l-glutamine, 50 U of penicillin/ml, 50 μg of streptomycin/ml, and 10% fetal calf serum, and used between passages 3 and 10 (42). Human DP cells were obtained from healthy human premolars as described by Wang et al. (48) and grown in RPMI 1640 as described for PDL cells. Between passages 3 and 10, the cells were grown to confluence in 96-well plates (5 to 7 days) to a final concentration of 5 × 10⁴ per well.

For some activation studies, THP-1, DP, or PDL cells were preincubated with vitamin D₃ (40 ng/ml) for 72 h. Prior to binding or activation experiments, cells were washed three times with serum-free RPMI 1640, incubated at 37°C for 24 h with serum-free RPMI 1640, and washed again with serum-free RPMI 1640. The number and viability of cells were examined by the 3-(4,5-dimethythiazol-2-yl)-2,5-diphenyltetrazolium test as described by Mosmann (29). Cell membrane components were extracted from pellets of NHS-LC biotin-labeled cells (10⁷ THP-1 cells; 2.5 × 10⁶ DP or PDL cells) (49) on ice with 0.2 ml of phosphate-buffered saline (PBS) containing 1% Nonidet P-40, 1 mM MgCl₂, 0.1 U of aprotinin/ml, and 2 mM phenylmethylsufonyl fluoride (34).

Binding assays.

Binding of biotinylated cell membrane extracts from THP-1, DP, or PDL cells to S. mutans cells or to purified protein I/IIf, RGP, or LTA was carried out as follows. Microtiter plates coated with either 50 μl of wild-type or isogenic mutant strains (2 × 10⁸ bacteria per ml) or purified protein I/IIf (10 μg/ml), RGP (20 μg/ml), or LTA (20 μg/ml) in 0.1 M carbonate buffer (pH 9.6) were first incubated with 200 μl of PBS-Tween containing 0.5% fish gelatin for 2 h at room temperature. After a wash, the biotinylated THP-1, DP, or PDL cell membrane extract (0 to 20 μg/ml in PBS-Tween) was added and incubated at 25°C for 1 h. After a wash, bound ligand was detected by sequential incubations with AP-streptavidin (for 1 h at 25°C) followed by enzyme substrate (for 1 h at 25°C) and reading of the optical density at 405 nm. Nonspecific binding was assayed in the presence of a 20-fold excess of unlabeled cell membrane extracts. The level of specific binding was calculated by subtracting nonspecific binding from total binding.

Activation of cells.

For activation experiments, cells were grown in 96-well plates as described above. THP-1 cells (2 × 10⁵ per well), DP cells (5 × 10⁴ per well), and PDL cells (5 × 10⁴ per well) were incubated at 37°C with 200 μl of RPMI 1640 containing various concentrations of live, heat-killed, or NaN₃-treated bacteria (0 to 100 bacteria per cell) or protein I/IIf (10 μg/ml), RGP (50 μg/ml), or LTA (50 μg/ml) in 200 μl of serum-free RPMI 1640 medium in the presence of polymyxin B (2 μg/ml). After various incubation periods, culture supernatants were harvested, centrifuged (at 3,000 × g for 10 min), and used to estimate TNF-α, IL-6, IL-8, and IL-10 release by a heterologous two-site sandwich enzyme-linked immunosorbent assay according to the manufacturer's instructions. In some cases, cells were pretreated with cytochalasin D (at 2.5 μg/ml for 45 min) before addition of the various stimuli. In inhibition assays, a constant amount of antibodies (10 μg/ml) was added to cells simultaneously with the various stimuli. For experiments using immobilized ligands, either 10⁷ bacteria or 0.5 μg of protein I/IIf or 2.5 μg of RGP or LTA in coating buffer was added to 96-well plates. After drying at 37°C and treatment with 0.5% fish gelatin, 2 × 10⁵ THP-1 cells were added to the wells and cytokine levels were measured as above.

Analytical procedures.

Total hexose was assayed by the resorcinol-sulfuric acid method of Monsigny et al. (28), and protein was measured by using a dye reagent (Bio-Rad, Marnes la Coquette, France) according to the manufacturer's instructions. Proteins were separated by SDS-PAGE on 7.5% polyacrylamide gels according to Laemmli's method (22) and either stained with Coomassie brilliant blue or electroblotted onto nitrocellulose sheets by the procedure of Towbin et al. (40).

Statistical analyses.

Student's t test was used for statistical analysis, and P values of <0.05 were considered significant.

RESULTS

Binding of S. mutans strains to THP-1, DP, and PDL cells.

To initiate cytokine synthesis, S. mutans must adhere to host cells by means of different adhesins. Therefore, we investigated the roles of both protein I/II and RGP in bacterial adherence by using stable deletion mutants for protein I/II (KT6219-11), RGP (KT6219DR), or both (KT6219DR-11). We first tested the binding of biotinylated THP-1 cell membrane extracts to the S. mutans wild-type strain KT6219 and to the three isogenic mutants defective in expression of either protein I/II, RGP, or both. The binding of biotinylated THP-1 cell membrane extracts was tested over a range of 0 to 20 μg/ml (Fig. 2A). THP-1 cell membrane extracts bound specifically to the four immobilized S. mutans strains in a saturable fashion (saturation was achieved at 10 μg of cell membrane extract per ml). Binding to isogenic mutants KT6219-11, KT6219DR, and KT6219DR-11 was less efficient (reduced by 34, 22, and 69%, respectively), and the lowest binding activity was observed for the isogenic double mutant KT6219DR-11. These results show that both protein I/II and RGP function as adhesins mediating the adherence of S. mutans to monocytes. This finding was confirmed by the ability of biotinylated THP-1 cell membrane extracts to bind specifically to immobilized purified protein I/II and RGP in a ligand-dependent fashion (Fig. 2B).

FIG. 2.

Dose-response analysis of the specific binding of cell membrane extracts to S. mutans strains (A) or to purified adhesins (B). (A) Binding of increasing amounts of biotinylated THP-1 cell membrane extracts to a plate coated with S. mutans KT6219 (▪), KT6219-11 (•), KT6219DR (▴), or KT6219DR-11 (○). (B) Binding of increasing amounts of biotinylated THP-1 cell membrane extracts to a plate coated with protein I/II (•), RGP (▴), or LTA (○). Specific binding was calculated for either 5 × 10⁶ bacteria, 0.5 μg of protein I/IIf, or 1 μg of RGP or LTA as the difference between total binding in the absence of unlabeled cell extracts and nonspecific binding in the presence of a 20-fold excess of unlabeled cell extracts. Results are expressed as means of triplicate determinations from three different experiments. Error bars, standard deviations from the means. The results are statistically significant: *, P < 0.05; **, P < 0.01; ***, P < 0.001.

The fact that the isogenic mutant KT6219DR-11, which is defective in both protein I/II and RGP expression, still binds to monocytes constitutes evidence in favor of the existence of other S. mutans cell surface adhesins, and we hypothesized that LTA could be such a molecule. We showed that the wild-type S. mutans strain KT6219, the three isogenic mutants, and purified LTA reacted with rabbit anti-LTA antibodies in an enzyme-linked immunosorbent assay (data not shown), demonstrating the presence of LTA on the bacterial cell surface. In order to examine the potential role of LTA, we tested the ability of biotinylated THP-1 cell extracts to bind to purified S. mutans LTA. The results indicate that LTA is able to bind to monocyte cell extracts and that the binding of THP-1 cell membrane extracts to LTA is twofold lower than the binding to protein I/II or RGP, thus demonstrating the role of LTA as a bacterial adhesin enabling the adhesion of S. mutans to monocytes.

The pattern of binding of biotinylated DP or PDL cell membrane extracts to the four S. mutans strains was similar to that of biotinylated THP-1 cell membrane extracts (data not shown): the binding activity of the parent strain, KT6219, was the highest, while that of an isogenic mutant lacking protein I/II or RGP was less effective, and the signal obtained with the mutant lacking both components, KT6219DR-11, was the lowest. Furthermore, we found that the binding of biotinylated DP and PDL cell membrane extracts to immobilized protein I/II, RGP, or LTA was effective and was similar to that observed previously with monocyte cell membrane extracts (data not shown). Taken together, our results demonstrate that both protein I/II and RGP play an important role in streptococcal adherence to human cells, whereas LTA might be only a weaker adhesin.

Cytokine release by THP-1, DP, and PDL cells after stimulation with S. mutans cells or purified adhesins.

The specific binding of S. mutans cells to THP-1, DP, and PDL cell extracts mediated by protein I/II, RGP, and LTA led us to investigate the role of each adhesin in stimulating the release of pro- and anti-inflammatory cytokines. THP-1, DP, or PDL cells were first incubated for 24 h with S. mutans KT6219 cells at variable concentrations in RPMI 1640, and optimal production of IL-8 was obtained at a bacteria-to-cell ratio of 50 to 1. Then THP-1, DP, or PDL cells were incubated with the S. mutans parent strain KT6219 (50 bacteria/cell); the stimulus was maintained for 0 to 48 h, and TNF-α, IL-6, IL-8, and IL-10 production was measured in the culture supernatants. The release of TNF-α, IL-6, and IL-8 was time dependent in THP-1 cells (Fig. 3). The maximal release was obtained between 12 and 24 h, after which levels declined slowly. In DP and PDL cells stimulated with the S. mutans wild-type strain KT6219, a time-dependent increase in IL-6 and IL-8 release was also observed, with maximal production at 12 to 24 h, whereas TNF-α was not generated above control levels (Fig. 3). Furthermore, incubation of cells with S. mutans strain KT6219 did not enhance the secretion of IL-10 above constitutive levels in THP-1, DP, and PDL cells (Fig. 3). The results show that S. mutans adhesins on cell surfaces may exert immunomodulatory effects, such as induction of proinflammatory cytokine release by THP-1, DP, and PDL cells, which are associated with the generation of a Th1 response. However, cytokines synthesized by the cells could also be associated with the modulatory activity of released bacterial components. Therefore, in order to minimize the putative role of secreted streptococcal components, and because significant levels of IL-6 and IL-8 were detected as early as 4 h, we decided to study the active role of each adhesin in the up-regulation of cytokine synthesis by measuring the release of IL-8 by cells stimulated for 4 h with either wild-type or mutant S. mutans cells or purified adhesins.

FIG. 3.

Time-dependent production of TNF-α (hatched bars), IL-6 (open bars), IL-8 (solid bars), and IL-10 (stippled bars) by THP-1, DP, and PDL cells stimulated with the S. mutans wild-type strain KT6219 (50 bacteria/cell). Each assay was carried out in triplicate. Error bars, standard deviations. Two additional experiments gave results similar to those shown here. Asterisks indicate significant differences from results for the respective control cultures without stimulation: *, P < 0.05; **, P < 0.01; ***, P < 0.001.

The response after a 4-h incubation of THP-1 cells with the parent strain KT6219 was a significant IL-8 release (15,000 pg/ml), whereas stimulation of THP-1 cells with either mutant strain KT6219-11, defective in protein I/II, mutant strain KT6219DR, defective in RGP, or mutant strain KT6219DR-11, defective in both adhesins, was less effective, resulting in 20, 65, or 85% reduction in IL-8 release, respectively (Fig. 4A). On the other hand, purified protein I/II seemed to be more efficient than RGP in triggering THP-1 cells to release IL-8 (Fig. 4B), whereas LTA had little effect on IL-8 up-regulation. Identical results were obtained with DP (Fig. 4C and D) and PDL (data not shown) cells stimulated with either whole bacteria (Fig. 4C) or purified adhesins (Fig. 4D). However, the amount of IL-8 released by DP (Fig. 4C and D) or PDL (data not shown) cells was three times lower than that released by THP-1 cells. These results suggest (i) that even if protein I/II seems to be more efficient in its purified form in triggering cells to release IL-8, RGP seems to play a more important role in whole bacteria and (ii) that LTA plays only a minor role in the cell activation process. In THP-1 (Fig. 4A and B), DP (Fig. 4C and D), and PDL (data not shown) cells preincubated for 72 h with vitamin D₃, an agent known to stimulate the expression of CD14, IL-8 release was greatly enhanced after incubation with either wild-type S. mutans, the isogenic mutant KT6219-11 defective in protein I/II expression, or purified RGP. In contrast, vitamin D₃ induced only a minor enhancement of IL-8 release in DP or PDL cells stimulated with either the RGP mutant, the double mutant defective in both protein I/II and RGP, purified protein I/II, or purified LTA (Fig. 4). These results suggest that CD14 acts as a receptor allowing stimulation by RGP of these three kinds of cells in order to release IL-8.

FIG. 4.

Production of IL-8 by THP-1 cells (A and B) or DP cells (C and D) stimulated with either S. mutans KT6219 (KT), S. mutans KT6219-11 (KT-11), S. mutans KT6219DR (KT DR), or S. mutans KT6219DR-11 (KT DR-11) (A and C) or with purified protein I/II, RGP, or LTA (B and D). Solid bars, untreated ligand; right-hatched bars, THP-1 and DP cells pretreated with vitamin D₃ (40 ng/ml) for 72 h; open bars, THP-1 and DP cells pretreated with cytochalasin D (2.5 μg/ml) for 45 min; stippled bars, THP-1 and DP cells incubated with mouse anti-CD14 monoclonal antibodies (10 μg/ml); shaded bars, heat-inactivated bacterial cells; left-hatched bars, bacteria (A) or adhesin (B) applied to microtiter plates. Data are expressed as means ± standard deviations for triplicate cultures, and the results are representative of three different experiments. Asterisks indicate significant differences from results for the respective control cultures without stimulation: *, P < 0.05; **, P < 0.01; ***, P < 0.001.

We investigated the role of CD14 in IL-8 synthesis further by incubating cells simultaneously with mouse anti-CD14 monoclonal blocking antibodies. When THP-1 cells were incubated with TÜK4 monoclonal antibodies, 55% inhibition was obtained after stimulation with S. mutans KT6219 or KT6219-11, and 45% inhibition was obtained after stimulation with RGP, whereas TÜK4 monoclonal antibodies had little or no effect on THP-1 cells stimulated with KT6219DR, KT6219DR-11, or protein I/II (Fig. 4A and B). In contrast, incubation of THP-1 cells with control IgG2a antibodies had no inhibitory effect on IL-8 release. Incubation of DP (Fig. 4C and D) or PDL (data not shown) cells with TÜK4 also suppressed the release of IL-8 in cells stimulated with S. mutans KT6219, S. mutans KT6219-11, or RGP. Moreover, the cell response to S. mutans could be associated with internalization of bacterial cells or the purified ligand. To test this hypothesis, wild-type and isogenic mutant bacteria were either inactivated by heat (60°C for 30 min) or treated with NaN₃ (at 50 mM for 3 h), and the IL-8 response was examined. Nonviable S. mutans cells showed only a slightly weaker ability to induce IL-8 release from stimulated cells (Fig. 4A and C). Furthermore, up-regulation of IL-8 production by the wild-type and mutant cells or the purified adhesins (protein I/II, RGP, and LTA) was not affected by cytochalasin D treatment of THP-1 (Fig. 4A), DP (Fig. 4C), or PDL (data not shown) cells, demonstrating (i) that binding of S. mutans to host cells is sufficient to induce IL-8 secretion and (ii) that internalization of streptococcal ligands is not required for the IL-8 response.

We further investigated the importance of ligand internalization in the cytokine response by an additional approach. Microtiter plates coated with either wild-type S. mutans cells, insertionally inactivated mutant S. mutans cells, or purified protein I/II, RGP, or LTA, to serve as a stimulus for IL-8 release, were incubated with THP-1 cells. THP-1 cells bound to coated ligands (Fig. 4A and B) secreted only slightly smaller amounts of IL-8 than THP-1 cells stimulated with soluble ligands. Taken together, these results confirm that cellular uptake through an intact cytoskeleton or an endosomal pathway is not required for S. mutans cells or S. mutans soluble adhesins to trigger an IL-8 response.

DISCUSSION

S. mutans, the causative agent of dental caries, is also an important pathogen in various inflammatory disorders, and the resulting acute or chronic inflammation is dependent on the induction of the synthesis of proinflammatory cytokines. Recently, we and others showed that various oral streptococci can stimulate epithelial and endothelial cells, monocytes, and lymphocytes to induce cytokines (18, 36, 45). However, among the numerous streptococcal components, only protein I/II and RGP (12) from S. mutans, superantigen from S. mutans (26), and secreted superoxide dismutase and the 190-kDa protein from S. sanguis (6) have been shown to be able to induce the synthesis of inflammatory cytokines. In the present work, we investigated the role of S. mutans protein I/II and RGP in the inflammatory process that occurs in dental pulpitis. To define the functions of protein I/II and RGP separately on the cell surface, we constructed and characterized an S. mutans mutant with functional deletions of both the pac and rmlB genes. We then used it together with the previously described S. mutans mutants defective in protein I/II and RGP to define more precisely the adhesin and cytokine up-regulatory activity of each component toward monocytic, DP, and PDL cells. We showed that protein I/II is implicated in the binding of S. mutans to monocytic, DP, and PDL cells, since a mutant defective in protein I/II did not interact as well with cells, whereas the binding to THP-1, DP, and PDL cells was similar. These results are consistent with the findings of different studies, using various oral streptococcus I/II deletion mutants, which showed that members of the I/II family bind to a broad spectrum of host molecules including salivary glycoproteins (32), extracellular matrix proteins (23), α₅β₁ cellular integrins (5), and Actinomyces naeslundii surface components (13). Furthermore, we show that an RGP deletion mutant binds less effectively to human cells, and this finding allowed us to conclude that RGP is an important streptococcal component in bacterial adhesion. Since the double-deletion mutant (i) interacts less with human cells than single-deletion mutants and (ii) still shows little residual binding activity, it can be postulated that protein I/II and RGP are the two most important adhesins, that adherence occurs through different receptors, and that additional bacterial cell surface proteins, especially the key gram-positive cell surface component LTA, appear to play only a minor role in the cellular adhesion process. In this respect, we can hypothesize that invasion and propagation of S. mutans cells in DP is dependent on the adhesive properties of both protein I/II and RGP and that the persistence of bacterial cells could be attributed to the antiphagocytic activity of RGP (43).

In the second part of this study, we analyzed the cytokine responses of monocytic, DP, and PDL cells elicited by the interactions of S. mutans protein I/II, RGP, and LTA with their cognate receptors. We confirmed that protein I/II and RGP, but not LTA, were potent stimulators of TNF-α, IL-6, and IL-8 production in monocytic THP-1 cells (8, 34), and we showed that LTA is a less potent cytokine inducer. In contrast, the secretion of TNF-α was not up-regulated in DP or PDL cells after stimulation with either whole S. mutans cells or purified adhesins. Furthermore, DP and PDL cells exhibited only 30% of the maximal IL-6- and IL-8-stimulating activities of THP-1 cells. These results are in agreement with previous observations (3, 45) that showed that cells such as epithelial or endothelial cells have a more selective proinflammatory cytokine response profile, probably serving to limit the local inflammatory response after microbial exposure. Unlike the OspA lipoprotein from Borrelia burgdorferi, which is able to induce production of IL-10 in THP-1 cells (31), intact S. mutans or purified protein I/II, RGP, or LTA never stimulated THP-1, DP, or PDL cells to release the anti-inflammatory cytokine IL-10. These results reinforce the notion that either constitutive or environmental bacterial components, in inducing an anti- or proinflammatory cytokine response, play a pivotal role in controlling the initial process of bacterial colonization and/or infection of the host and the development of inflammation and immune responses. Our data also confirm earlier results from Jiang et al. (21), who showed that S. mutans induces the synthesis of high levels of type 1 cytokines which are associated with the generation of a Th1 response, in contrast to Lactobacillus casei, which preferentially induces a Th2 response (18), and they suggest that S. mutans antigens such as protein I/II and RGP could contribute to chronic pulpal inflammation which occurs under shallow lesions.

Although protein I/II and RGP seem to be equally important in bacterial adherence, and since purified proteins seem to be more-potent cell activators, RGP is the most efficient cytokine-stimulating component in intact bacteria. This could be attributed either (i) to the relative abundance of the receptors for RGP versus protein I/II, (ii) to the hydrophilic nature and flexibility of polysaccharides, which allow RGP to bind more avidly to its cognate receptor, partially masking the interactions of S. mutans with cells via protein I/II by steric hindrance, (iii) to the relative affinity of each component for its cognate receptor, or (iv) to the efficiency of the signal transduction pathway.

Next, by using either heat, NaN₃, solid-phase-adsorbed bacteria, or pretreatment of cells with cytochalasin D, we showed that the binding of S. mutans, purified protein I/II, or RGP to THP-1, DP, or PDL cells is sufficient to induce an IL-8 response in the absence of significant uptake of the ligands. These results are in agreement with previous data which showed that the binding of LPS to CD14, Toll-like receptors, and β₂ integrins is capable of triggering intracellular signals leading to cytokine synthesis in various cells in response to cell receptor occupancy (41). They are also consistent with studies that have shown that ligand binding to integrins (5) induces the clustering of integrins within the cell membrane, which activates intracellular signaling pathways such as tyrosine phosphorylation of intracellular proteins that coordinate a variety of cellular responses associated with cytoskeleton reorganization.

The fact that preincubation of THP-1, DP, and PDL cells with vitamin D₃, which is known to stimulate the expression of CD14 on the cell surface, greatly enhances the release of IL-8 by the cells after stimulation with wild-type S. mutans, the protein I/II deletion mutant, or RGP demonstrates that CD14 acts as a receptor for RGP, enabling the release of IL-8 in response to stimulation of THP-1, DP, and PDL cells, and that CD14 is constitutively expressed on the surfaces of all three types of cells. These results also indicate that the signaling pathway used by LPS or other bacterial lipoconstituents may be shared by RGP, and since RGP is present in the cell walls of oral streptococci and Lancefield group A, C, and E streptococci, RGP seems to be a major component in streptococcal virulence. On the other hand, the fact that the stimulating capabilities of the RGP deletion mutant, the double-deletion mutant, and protein I/II were not affected by preincubation of cells with vitamin D₃ confirms that CD14 does not act as a receptor for protein I/II. These results are also confirmed by preliminary data which show that incubation of THP-1 cells together with protein I/II and RGP produces a cooperative effect on the release of IL-8. Despite the fact that LTAs from different gram-positive bacteria exhibit a CD14-dependent ability to trigger various cells for cytokine release, we showed that the protein I/II deletion mutant, the double-deletion mutant, and purified LTA only slightly stimulate THP-1, DP, and PDL cells to release IL-8. However, these results are in agreement with other studies showing a great variability in LTA structure (16), which could be responsible for the inability of LTA to induce cytokine synthesis in some cases (38, 47).

In conclusion, we demonstrated that the binding of S. mutans to human monocytic, DP, and PDL cells is mediated principally by protein I/II and RGP and that the other bacterial cell surface components, including LTA, are only minimally implicated in binding to host cells. These results are consistent with those of Bolken et al. (9), who showed that inactivation of the srtA gene in Streptococcus gordonii inhibits cell anchoring of protein I/II and decreases the in vivo adhesion of S. gordonii to mouse oral mucosal surfaces. Furthermore, we showed that the binding of protein I/II and RGP to their cognate receptors on cells accounts for the major proinflammatory activity of S. mutans cells and that internalization of both ligands is not necessary to trigger cells to release IL-8. We also showed that, beside the implication of monocytes in pulpal inflammation, fibroblast-like cells such as DP and PDL cells are actively implicated in the local inflammation initiated by S. mutans cells. The present findings confirm and suggest new therapeutic possibilities for the treatment of endodontic inflammation. The administration of either RGP or protein I/II antagonists could block the cell activation process due to S. mutans and thus reduce local pulpal activation.