|Home | About | Journals | Submit | Contact Us | Français|
The oral mucosal pathogen Porphyromonas gingivalis expresses at least two adhesins: the 67 kDa mfa-1 (minor) fimbriae and the 41 kDa fimA (major) fimbriae. In periodontal disease, P. gingivalis associates in situ with dermal dendritic cells (DCs), many of which express DC-SIGN (CD209). The cellular receptors present on DCs that are involved in the uptake of minor/major fimbriated P. gingivalis, along with the effector immune response induced, are presently unclear. In this study, stably transfected human DC-SIGN+/- Raji cell lines and monocyte-derived DCs (MoDCs) were pulsed with whole, live wild-type Pg381, isogenic major- (DPG-3), minor- (MFI) or double-fimbriae (MFB) deficient mutant P. gingivalis strains. The influence of blocking antibodies, carbohydrates, full-length glycosylated HIV-1 gp120 envelope protein and cytochalasin D on uptake of strains and on the immune responses was determined in vitro. We show that binding of minor fimbriated P. gingivalis strains to Raji cells and MoDCs is dependent on DC-SIGN, while the double-fimbriae mutant strain does not bind. Binding to DC-SIGN on MoDCs is followed by internalization of P. gingivalis into DC-SIGN rich intracellular compartments and MoDCs secrete low levels of inflammatory cytokines and remain relatively immature. Blocking DC-SIGN with HIV-1 gp120 prevents uptake of minor fimbriated strains and deregulates expression of inflammatory cytokines. Moreover, MoDCs promote a Th2 or Th1 effector response, depending on whether they are pulsed with minor or major fimbriated P. gingivalis strains, respectively, suggesting distinct immunomodulatory roles for the two adhesins of P. gingivalis.
The C-type lectin DC-specific ICAM-3 grabbing non-integrin (DC-SIGN) is a pattern recognition receptor (PRR) and adhesion molecule expressed by dendritic cells (DCs) and by certain types of macrophages (1). It is used to endocytose microbial antigens in the periphery, to bind to ICAM-2 on endothelial cells (2) and to mediate immune clustering with ICAM-3+ T cells in the lymph nodes (1, 3, 4). It is also expressed on blood DCs (5) and in pathological conditions such as rheumatoid arthritis (6) and in rupture prone atherosclerotic plaques (7, 8). Recent studies indicate that DC-SIGN+ DCs increase in the oral mucosal disease chronic periodontitis (CP) (9-11). DC-SIGN is one of a family of calcium-dependent C-type lectins that bind to carbohydrate motifs and to Lewis blood group antigens (12, 13). Although lacking Toll-IL-1r activation domains, DC-SIGN has emerged as a key player in the induction of immune responses against numerous pathogens, via modulation of TLR-induced immune activation (14). This occurs by activation of ERK (15), and Raf-1 kinase-dependent acetylation of transcription factor NF-κB (16, 17). Mycobacterium tuberculosis (18) Mycobacterium leprae (19), and Helicobacter pylori (20) target DC-SIGN to gain entry into DCs, disrupt full DC maturation and inhibit Th1 effector cell polarization. Neisseria meningitidis and Lactobacillus spp., on the other hand, target DC-SIGN to modulate the immune response towards Th1 (21) or Treg (22), respectively.
The immunopathogenesis of chronic periodontitis (CP) has been linked to negative regulation of TLRs (23-25) and to the presence of Th2 effector T cell populations (reviewed in (26)), but the specific role of oral mucosal pathogens in induction of Th2 effector responses are just beginning to be identified (9). The oral mucosa in CP contains organized lymphoid aggregates, called oral lymphoid foci, or OLF (27). OLF contain immune conjugates consisting of dermal DCs and CD4+ T cells, as well as B cells (28). Of particular interest is the presence of an intense infiltrate of DC-SIGN+ DCs in the lamina propria of CP, combined with evidence that DCs in the lesions appear to mobilize towards the capillaries (28). This has fueled speculation that, as with gut lamina propria DCs (29), specific microbiota in the oral mucosa target lamina propria DCs that can direct the T cell effector responses (30, 31).
P. gingivalis is one of several intracellular pathogens implicated in CP (reviewed in (32)). Most pathogens, P. gingivalis included (33) express different pathogen-associated molecular patterns (PAMPs) that can trigger distinct classes of PRRs on a single cell simultaneously (14). Of particular relevance are the two adhesins of P. gingivalis, termed the mfa-1 (minor) and fimA (major) fimbriae. Adhesion of pathogens to host tissues and subsequent invasion are important early events in mucosal pathogenesis (34). The minor and major fimbriae of P. gingivalis have been shown in the rat model to play roles in the pathogenesis of periodontal disease (34). The two fimbriae are distinct antigenically, by amino acid composition, and by size (35, 36). The major fimbriae is composed of a 41 kDa protein, encoded by the fimA gene (37). Much is known of the PRRs targeted by the major fimbriae (38-42) and of the intracellular signaling pathways that are activated (43, 44). In contrast, little is known of the cellular receptors targeted by the 67 kDa minor fimbriae, encoded by the mfa1 gene. Expression of both fimbriae is regulated under different environmental conditions (45-47) Understanding the immunobiological properties of these two fimbriae could help in understanding how this oral mucosal pathogen evades the immune response and induces periodontal disease, described as a Th2 type disease (24).
The purposes of the present study were: (i) to determine the role of DC-SIGN in binding and uptake of isogenic minor and major fimbriae-deficient mutants of P. gingivalis using stably transfected Raji (B-) cell lines and monocyte-derived dendritic cells (MoDCs), and; (ii) to determine how minor/major fimbriae influence DC maturation, cytokine secretion and the T cell effector responses induced by MoDCs. Our results show that the minor fimbriae of P. gingivalis are required for binding to the endocytic receptor DC-SIGN, leading to internalization in DC-SIGN rich compartments. This uncouples cytokine secretion from maturation of DCs and elicits a Th2-biased effector T cell response. Overall these results may help explain how this oral pathogen evades and suppresses the immune response.
WT Pg381, which expresses both minor and major fimbriae (Pg min+/maj+), isogenic minor fimbriae-deficient mutant MFI, which expresses only the major fimbriae (Pg min-/maj+), isogenic, major fimbriae-deficient mutant DPG3, which expresses only the minor fimbriae (Pg min+/maj-), and the double fimbriae mutant MFB (Pg min-/maj-) were maintained anaerobically (10% H2, 10% CO2, 80% N2) in a Forma Scientific Anaerobic System glove box model 1025/1029 at 37°C (48, 49) in Difco Anaerobe Broth MIC. Erythromycin (5 μg/ml) and tetracycline (2 μg/ml) were added according to the selection requirements of the strains. Bacteria were pelleted, washed once in phosphate buffered saline (PBS) and for FACS based analyses, stained with CFSE (Molecular Probes, Eugene Oregon, USA), as described (50). Briefly, bacteria in PBS were stained with CFSE at a final concentration of 5μM for 30 min at 37°C and protected from light. The bacterial suspension was washed five times in PBS and P. gingivalis were resuspended to an O.D. at 660nm of 0.11, previously determined to be equal to 5 × 107 CFU (51). MoDCs and Raji cells were pulsed with P. gingivalis strains at 5:1 or 25:1 multiplicity of infection (MOI) for from 1.5 to 18 hr. Low MOI's were used to mimic a natural infection as well as to avoid overwhelming the host response. The percentage viable MoDC or Raji (typically >90% after 18 hrs) were monitored by trypan blue exclusion and did not differ between the strains (not shown). MoDCs and Raji cells were fixed and stained for FACS analysis of %Pg-CFSE+ cells as described in figure legends.
Raji cell lines (52) were obtained courtesy of D. R. Littman (Skirball Institute of Biomolecular Medicine, NYU), maintained in 10% heat inactivated FBS (GIBCO), RPMI 1640 with L-Glutamine and NaHCO3 (SIGMA) in a 5% CO2 incubator at 37°C. Staining to verify surface receptor expression was performed using monoclonal antibodies (BD-Biosciences): FITC anti-CD14 (cat # 555397); anti-CD209 (cat # 551264); anti-E-Cadherin (cat # 612130); anti-CD19 (cat #557697); IgG1 isotype (cat # 349041); anti-CD80 (cat# 557226); IgG2a isotype (cat # 349051); PE anti-CD29 (cat# 555443); anti-CD209 (cat# 551265); anti-CD18 (cat# 555924); anti-CD206 (cat# 555954); anti-HLA-DR (cat# 555812); anti-CD86 (cat# 555657); anti-CD11a (cat# 555379); IgG1 isotype (cat #349043); APC anti-CD205 (cat# 558156); anti-CD11b (cat# 550019); IgG1 isotype cat# 557732; (eBiosciences) PE anti-human TLR4 (cat# 12-9917-73); FITC anti-human TLR4 (cat# 53-9917-73); FITC anti-human TLR2 (cat# 11-9922-73); (Immunotech) FITC CD83 (cat# PN IM2410); (Immunotech) PE anti-CD207 (cat# IM3577); (Invitrogen) PE anti-CD11c (cat# 349863a); (Dako) FITC anti-CD1a (cat# F7141).
CFSE- or Syto-stained bacteria were added at an MOI of 5:1 to either Raji or Raji DC-SIGN (52, 53). Binding was conducted on ice to prevent loss of surface DC-SIGN, which cycles rapidly on Raji cells at 37°C (52, 53, 93). Cells were pulsed with CFSE stained bacteria for from 1-18 hr either at 37°C, on ice (48) or in the presence of cytochalasin D (0.5 μM the minimal concentration needed to arrest cytoskeletal rearrangements in Raji cells). Cells were washed to remove unbound bacteria, fixed in 1% formalin and analyzed by FACS. Association was quantified via FACS as previously described (19, 54, 55). In brief, 10,000 Raji cells were gated on forward and side scatter characteristics based on size and to exclude debris and unbound bacteria and then %CFSE positive cells in FITC channel quantitated.
Carbohydrates were purchased from SIGMA: D-Mannose (cat# M-6020); L-Fucose (cat# F2252-5G); D-Fucose (cat# F-8150); Mannan from Saccharomyces cerevisiae (cat# M7504-5g) were all diluted into 2% heat inactivated FBS (GIBCO) PBS and filter sterilized. 5-100 μg of carbohydrates were added to block DC-SIGN receptor. The following reagent was obtained through the NIH AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH: HIV-1 gp120 CM envelope protein (Cat #2968). 1.5-9 μg of gp120 was added to block DC-SIGN receptor. Blocking antibodies to human DC-SIGN were purchased from R&D systems (cat# MAB161) (clone 120507); IgG2B Isotype control antibodies (cat#MAB004) (clone 20116); Anti-human integrin beta2 (CD18) mAb from Chemicon (cat#CBL158); Anti-human integrin beta1 mAb from Chemicon (cat#MAB1987Z); Anti-human CD11c mAb from BD Pharmingen™ (cat#555390 clone B-ly6); IgG1 isotype control from Chemicon (cat#CBL600) were added at a final concentration of 10 μg/ml. Raji and Raji DC-SIGN cells were pre-incubated with blocking antibodies or carbohydrates for at least 30 min on ice. CFSE stained bacteria were then co-cultured for 1 hr on ice and binding was measured via FACS as described in adhesion assay. MoDC's were pre-incubated with carbohydrates, HIV-1 gp120 CM envelope protein, or antibodies for at least 30 min at 37°C before being co-cultured with CFSE stained bacteria. Co-cultures proceeded for 3 hr at 37°C and association was measured via FACS as described in the Raji adhesion assay.
MoDCs were generated as we have previously described (10, 11, 27). Briefly, monocytes were isolated from mononuclear fractions of peripheral blood of healthy donors and seeded in the presence of GM-CSF (100 ng/mL) and IL-4 (25 ng/mL) at a concentration of 1–2 × 105 cells/mL for 6–8 days, after which flow cytometry was performed to confirm the immature DC phenotype (CD14- CD83- CD1a+DC-SIGN+). Cell surface markers of DCs were evaluated by four-color immunofluorescence staining with the following mAbs: CD1a- FITC (Biosource), CD80-PE (Becton Dickinson), CD83-PE (Immunotech), CD86- PE (Pharmingen), HLA-DR-PerCP (Becton Dickinson), CD14- APC (Caltag). After 30 minutes at 4°C and washing with staining buffer (PBS, ph7.2, 2 mM EDTA, 2%FBS), cells were fixed in 1% paraformaldehyde. Analysis was performed with FACScalibur™ (Becton Dickinson). Marker expression was analyzed as the percentage of positive cells in the relevant population defined by forward scatter and side scatter characteristics. Expression levels were evaluated by assessing mean fluorescence intensity (MFI) indices calculated by relating MFI noted with Pg-CFSE or the relevant mAb to that with the isotype control mAb for samples labeled in parallel and acquired using the same setting.
Culture supernatants were collected from MoDC's pulsed with P. gingivalis strains for 3 and 18 hr. Culture supernatants were analyzed by flow cytometry using a cytometric bead array (CBA kit; BD Biosciences, San Diego, Calif.) following manufacturers' instructions. A standard curve was achieved for each cytokine, the CBA software calculates levels in picograms per milliliter.
For T cell priming experiments, the responder cells were autologous CD4+ naïve T cells purified from mononuclear fraction of human buffy coats by positive selection, using anti-CD4 MAb and goat anti-mouse immunoglobulin G-coated microbeads (Miltenyi Biotech GmbH, Gladbach, Germany) as described previously (10). Briefly, isolation of CD4+ cells was achieved using Minimacs separation columns (Miltenyi Biotech GmbH) as described by the manufacturer. In all experiments the isolated cells were 80 to 90% CD4+, as determined by staining with fluorescein isothiocyanate-conjugated anti-CD4 MAb followed by flow cytometry analysis (results not shown). MoDCs were washed extensively after an 18-hr pulsing with P. gingivalis strains and cultured at graded doses (1,000, 500, and 50 DCs, all per 200 μl) in complete RPMI medium with 10% heat-treated fetal calf serum with autologous T cells (50,000 cells/ 200 μl). Proliferation was determined after 5 days by loss of CFSE staining.
P. gingivalis has been previously shown to associate in situ with DCs in human oral mucosa from chronic periodontitis patients (56) and to be taken up by MoDCs in vitro (57) but the cellular receptors used for binding to MoDCs and uptake are unclear. DC-SIGN is of particular interest as DC-SIGN+ DCs increase in inflamed oral mucosa in chronic periodontitis (28, 58). To determine the ability of wt Pg381 to bind to DC-SIGN, stably transfected DC-SIGN positive (Raji-DCS) and negative Raji cells (Raji) were obtained and the phenotype verified by flow cytometry (Fig. 1A). Our results indicate that Raji-DCS and Raji are positive and negative, respectively, for DC-SIGN, as previously reported (52). The expression level of other relevant cell surface receptors are also shown in Fig 1A. To determine binding of wt Pg381 (Pg min+/maj+) to Raji-DCS and Raji, bacteria were syto- or CFSE-labeled and binding analyzed qualitatively by image-enhanced fluorescence microscopy at low magnification (Fig. 1B, panels 1-4) and at higher magnification (Fig. 1B, panels 5, 6). Our results indicate that DC-SIGN is required for optimum binding of wt Pg381 to Raji cell lines. This disparity in binding to Raji-DCS vs. Raji was quantitated by flow cytometry, with binding analyzed at 1.5 hr, 3 hr and 6 hr. Optimum binding was achieved at 1.5 hr. Shown in Fig. 1C is the percentage of Raji-DCs and Raji that were associated with wt Pg381 at 1.5 hr. The results indicate a significantly decreased binding of wt Pg381 to Raji vs. Raji-DCS (↓43%, p<0.05, Student t-test) (Fig. 1C, Fig 1D). This decrease in binding to Raji was greater (↓71%) with strain DPG-3. In contrast, there was no difference in binding of MFI or MFB to Raji vs Raji-DCS. Binding of MFI to Raji likely depends on CD29, expressed equally by Raji-DCS and Raji (Fig 1A) and previously shown to bind major fimbriae (38). MFB failed to associate with either Raji-DCS or Raji (Fig 1D). To further verify the role for DC-SIGN in binding of minor vs major fimbriated P. gingivalis strains to Raji-DCS, we blocked DC-SIGN with L-fucose, mannose and mannans (Fig. 2). D-fucose, a stereoisomer of L-fucose, does not block DC-SIGN (59) was used as a negative control for sugar blocking. Preliminary studies established optimum dose (50 μg/ml) of sugars for blocking of DC-SIGN on Raji DCS by FACS analysis (not shown). We show that binding of wt Pg381 and DPG-3 to Raji-DCS was blocked by L-fucose, D-mannose and mannans, but not control D-fucose (Fig. 2). Strain MFI was not blocked by any of the sugars; nor did the sugars result in significant blocking of any of the strains to Raji cells (not shown). Blocking of MFB with sugars was not performed as this strain did not bind to either Raji cell line.
Phenotypic analysis of day 6 MoDCs by flow cytometry indicates that MoDCs express surface DC-SIGN, as well as CD29, CD11b, CD11c, CD18 and DEC-205 (Fig. 3A). Human MoDCs were pulsed with the four CFSE-labeled strains. The results of flow cytometry analysis (Fig. 3B) indicate that association of three of the P. gingivalis strains with MoDCs occurred within 3 hr, in the following order: MFI > wt Pg381 > DPG-3. MFB did not bind to MoDCs over background of 3%. Antibody blocking studies were thus performed with all strains except MFB. Shown in Fig. 4A are results with wt Pg381. Antibody blocking studies revealed that anti-DC-SIGN, but not anti-CD11c, anti-CD18 or anti-CD29 resulted in a significant reduction in association of wt Pg381 with MoDCs. Endocytosis of FITC-dextran (60) was unaffected by anti-DC-SIGN antibody (not shown), indicating that phagocytosis was still intact. Use of cytochalasin D, which inhibits actin polymerization required for internalization, but not binding, demonstrates that P. gingivalis is being internalized by MoDCs. DC-SIGN-blocking sugars mannose and mannan diminished uptake of wt Pg381 (Fig. 4A) (13). As mannose is a minor component sugar of the LPS of P. gingivalis (61, 62) we tested the ability of P. gingivalis LPS to block uptake of wt Pg381, but there was no effect (data not shown). To further confirm the role of DC-SIGN on MoDCs in binding to P. gingivalis, we used DC-SIGN-targeting HIV-1 glycosylated envelope protein gp120 as a blocking agent (Fig. 4B). We show that gp120 resulted in a dose-dependent loss of uptake of wt Pg381 and DPG-3, but not MFI to MoDCs.
To visualize extracellular and intracellular association of syto-labeled wt Pg381, MoDCs were probed with FITC-labeled anti-DC-SIGN at 1 hr (Fig. 5A) and 6 hr (Fig 5B, Fig 5C), then analyzed by image enhanced fluorescence microscopy, aided by deconvolution analysis. Early attachment to surface DC-SIGN (Fig 5A, arrows) is followed by intracellular localization of P. gingivalis with DC-SIGN in MoDCs (Fig 5B, 5C, arrows). At later time points (18 hrs), large numbers of essentially intact wt Pg381 were detected within as yet undefined intracellular compartments (Pg-containing vesicles or PgCV) of MoDCs (Fig 5D).
Microbial DC-SIGN ligands reportedly dampen TLR-dependent production of inflammatory and Th1-biasing cytokines by MoDCs (20). We therefore analyzed the production of inflammatory cytokines by MoDCs pulsed with the P. gingivalis strains for 3 hr (Fig. 6A) and 18 hr (Fig. 6B). We show that MFI was the most potent inducer of IL-1β, IL-8, IL-6 and TNFα at 3 hr and of IL-1β, IL-10, IL-12p70, IL-8 and TNFα at 18 hr. In contrast, DC-SIGN targeting strain DPG-3 induced no detectable TNFα and significantly lower levels of IL-1β (vs. wt Pg381 or MFI) at 3 hr and of IL-1β, IL-12p70, IL-8, IL-6 and TNFα (vs. MFI) at 18 hr. Double mutant MFB induced the lowest levels of nearly all cytokines at 3 hr and 18 hr. To further confirm the influence of DC-SIGN on regulation of inflammatory cytokine production by P. gingivalis, DC-SIGN was blocked with HIV-1 gp120 prior to pulsing with all strains except the double mutant. The results for wt Pg381 (Fig 6C) indicate that blocking DC-SIGN enhances the induction of inflammatory cytokines IL-1β, IL-12p70, IL-8 and IL-6, but not IL-10. In contrast, in the absence of the activating major fimbriae (i.e. DPG-3), blocking DC-SIGN with gp120 does not increase cytokines IL-12p70, TNFα and IL-6. Interestingly, HIV-1 gp120 blocking of DC-SIGN also deregulated inflammatory cytokine production by MFI, which does not express DC-SIGN ligand. HIV-1 gp120 alone did not induce inflammatory cytokines, but did induce IL-10, as previously reported (63).
DCs were further analyzed for maturation status at 18 hr (Fig 7A-C). We show that, while HLA-DR induction was nearly equivalent for all strains, strain MFI was the strongest inducer of CD80 and CD83, and MFB was the weakest inducer of all co-stimulatory and maturation markers. Compared to MFI, strain DPG-3 was a relatively weak inducer of CD80, CD83 and CD86, with wt Pg381 falling somewhere between DPG-3 and MFI (i.e. in induction of CD80 and CD86). Blocking DC-SIGN with HIV-1 gp120 inhibited MoDC maturation as was previously published by Shan et al. 2007 (63). Furthermore, pre-incubation with HIV-1 gp120 prior to co-culture with P. gingivalis strains inhibited the MoDC maturation induced by DC-SIGN targeting strains wt Pg381 and DPG-3, unlike the inflammatory cytokine response, which was enhanced by HIV-1 gp120 (Fig 6C). The presence of HIV1 gp120 inhibited DPG-3-induced co-stimulatory molecules CD80, CD86 and CD83 upregulation the strongest, followed by wt Pg381. Blocking DC-SIGN did not block upregulation of HLA-DR, CD80 or CD83 induced by non-DC-SIGN targeting strain MFI (Fig 7B,C). These results suggest an uncoupling of DC maturation from the inflammatory cytokine response when DCs phagocytose whole live bacteria that express a DC-SIGN ligand.
Previous studies have shown that DC-SIGN ligands can induce a Th2-based effector response (20, 63). In the present study, MoDCs were pulsed with each of the four P. gingivalis strains, then co-cultured with autologous naïve CD4+ T cells for 7 days, after which T cell cytokines and T cell proliferation were analyzed. We show that MoDCs pulsed with DC-SIGN-binding strain DPG-3 induced release of significantly higher IL-4 levels from T cells compared to all other strains and very low levels of IL-12p70 relative to MFI (Fig. 8A). When expressed as Th1/Th2 cytokine ratios (IFNγ/IL-4 and IL-12p70/IL-4), DPG-3-pulsed MoDCs induced the lowest levels of all strains except for MFB and yielded a very low Th1 index (Table). In contrast, MFI pulsed MoDCs induced significantly lower levels of the Th2 cytokine IL-4, but comparable levels of IFNγ and very high levels of IL-12p70 from T cells. This yielded high ratios of IFNγ/IL-4 and IL-12p70/IL-4 and a relatively high Th1 index (=8.81) (20). Wt Pg381 induced the lowest levels of IL-4 low levels of IL-12p70, and comparable levels of IFN-γ by T cells. The low levels of IL-4 by MFI resulted in the highest Th1 index (10.63) while DPG-3 induced the lowest levels of IFN-γ and IL-12p70 and had the lowest Th1 index (=1.15) (Table).
To determine if the immuno-proliferative ability of P. gingivalis pulsed MoDCs would correlate with the Th1 index, CFSE-labeled naïve CD4+ T cells were analyzed at various stimulator: effector ratios for 7 days. 1:50 MoDC-T cell ratio yielded the maximum proliferation. Our results (Fig 8B) show that MoDCs pulsed with DPG-3 and MFB induced the weakest T cell proliferative responses, while wt Pg381 and MFI induced the strongest T cell proliferative responses. The weak T cell proliferation and cytokine secretion exhibited by the double fimbriae knockout MFB could be attributed to its lack of binding and uptake (Fig 3B). However, we can attribute the impaired proliferation of co-cultured CD4+ T cells to the minor fimbriae of DPG-3 interacting with DC-SIGN. Furthermore, we can attribute the robust IL-12p70 cytokine production of MFI to that strain's lack of immunosuppressive DC-SIGN targeting fimbriae. Linear regression analysis of % T cell proliferation induced by MoDCs pulsed with each strain, showed a significant association (r2 = 0.857, p=.024 [SPSS, ver 15]) with the Th1 index.
Overall, these results indicate that both the major and minor fimbriae of P. gingivalis are involved in binding of the whole live bacterium to Raji cell lines (Fig 1) and to DCs (Fig 3); however, the minor fimbriae is required for binding to DC-SIGN (Figs 1, ,2,2, ,4).4). This results in P. gingivalis being internalized and routed in large numbers into as yet undefined intracellular vesicles of DCs (Fig. 5). DCs that have internalized P. gingivalis strains that lack the major fimbriae are poorly matured (Fig. 7), secrete very low levels of inflammatory cytokines (Fig 6) and induce a Th2-biased, weak immunoproliferative T cell response (Fig 8). While these findings were established using isogenic fimbriae-deficient mutants of P. gingivalis, expression of the fimbriae have been shown to be regulated by growth conditions, including temperature (45, 64) and hemin levels (46) Apparently, a two-component regulatory system (FimS/FimR) controls the two fimbrial genes at different levels depending on heme and temperature (47). Other systemic mucosal pathogens such as Yersinia pseudotuberculosis (65), Salmonella enterica (66, 67) also regulate their invasive potential under environmental pressures. This is of particular relevance here since the preferred ecological niche of P. gingivalis, a hemin-requiring anaerobe are deep bleeding “crypts” in the human oral mucosa, called periodontal pockets (31). These pockets are subjacent to OLF and to lamina propria dermal DCs (27) where P. gingivalis infects DCs in situ (56). P. gingivalis initially colonizes surface mucosa and tooth surfaces, where hemin levels and temperature are reduced, forming part of a complex biofilm (31, 68). The major fimbriae appear required for initial attachment to host cells (48, 57, 69-71), while the minor fimbriae appear to play an important role in microcolony formation by facilitating cell–cell interactions and promoting biofilm formation (68, 72, 73). Both fimbriae play essential roles in induction of alveolar bone loss (34) and atherosclerosis (74) in rats, but the specific mechanisms of these (seemingly) disparate processes are unclear. A recent review describes the important role for Th2 type responses in the inability of the host to successfully resolve periodontal disease (26), suggesting clinical relevance to our findings of Th2-responses biased by the minor fimbriae of P. gingivalis.
Relatively little is known about the steps involved in the formation and secretion of minor fimbriae and of the cellular receptors and signaling pathways it targets. In contrast, the major fimbriae is under intense investigation in this regard (31, 35, 36, 39-41, 42). The major fimbriae activate macrophages through TLR2 and TLR4, as well as complement receptor 3 and CD14 (43, 44, 75). Davey et al. (2008) showed that both the major and minor fimbriae specifically bind to chimeric TLR2 and CD14 proteins in a cell free ELISA (49). TLR2 appears to be particularly important in IL-10-mediated mucosal immune homeostasis in response to commensals (76). DC-SIGN-ligation (using purified microbial ligands) has also been shown to induce IL-10, by triggering Raf-1 phosphorylation (16). Combined with activation of TLR4, DC-SIGN ligation results in enhanced and prolonged NFκB activation and stronger IL-10 production (16). Our cytokine results corroborate cross-talk between DC-SIGN and TLRs. For example, the most profound effect that blocking DC-SIGN had on cytokine secretion was observed with wt Pg381 (Fig 6C), which expresses ligands for DC-SIGN (minor fimbriae) and TLR2/4 (major fimbriae). With Pg381 we saw an enhancement (deregulation) of IL-1β, IL-12p70, IL-8 and IL-6. In contrast, blocking DC-SIGN did not enhance IL-1β, IL-12p70, or IL-6 in response to strain MFI, which lacks DC-SIGN ligand minor fimbriae. Moreover, in the absence of major fimbriae (DPG-3), blocking DC-SIGN decreased IL-12p70 and IL-6. TNFα, a good indicator of NFkβ activation was consistently dampened by gp120 (Fig 6C), regardless of bacterial strain used. Although DC maturation and cytokine secretion are both generally attributed to ligation of TLRs leading to NFκβ activation (77), phagocytosis/endocytosis in itself triggers a family of intracellular signaling pathways (reviewed in (78-80). Consistent with this concept, we show that blocking DC-SIGN reduces phagocytosis of P. gingivalis by MoDCs (Fig 4), and reduces upregulation of costimulatory molecules (Fig 7B, 7C) but not certain cytokines (Fig 6). Overall, these results suggest an uncoupling between the DC cytokine response and maturation by DC-SIGN ligands that warrants further analysis with purified native fimbriae.
In this context, we recently purified the minor fimbriae by HPLC and analyzed the protein sequence by MALDI-TOF. We discovered that there are two conserved Asn-Xaa-Ser/Thr N-glycosylation motifs on the minor fimbriae sequence (81). Putative glycosylation of the 67 kDa minor fimbriae was further verified by Pro-Q Emerald glycoprotein staining, a periodate- and fluorescence-based reaction (unpublished data). Efforts are underway to analyze the purified minor fimbriae by monosaccharide compositional analysis and LC-MS/MS. There are other reports of a role for glycosylation of the fimbriae in P. gingivalis. Knockouts of gftA (a wcaE glycotransferase homolog of E. coli) in P. gingivalis fail to make mature fimbriae (82). The gingipains of P. gingivalis are apparently glycosylated (83) and this glycosylation is regulated by the vimF, vimA and vimE glycotransferase genes (84, 85). Knocking out these genes causes a failure to glycosylate these gingipains, leading to their inactivation (83-85). Kadowaki et al. (1998) have identified that Arg gingipain activity is essential for the processing and translocation of mature fimbriae (86). Recently, it was discovered that the RagA proteins of P. gingivalis are glycosylated (87). The commonalties among these outer membrane proteins is that they encode for an N-terminal, long signal peptide that gets cleaved once they enter the inner membrane (88-90). Many mucosal pathogens exhibit glycosylation motifs on their flagella, pili, and fimbriae for binding to host cells (91). Glycosylation also reportedly plays a role in maintaining the protein structure, in protection of proteolytic degradation and in immune evasion (81, 91). Additionally, it was recently determined that the glycosylation of soluble peanut allergen was sufficient for targeting DC-SIGN in MoDCs and that its recognition by DC-SIGN skewed the T cell response to a Th2 phenotype (92). Finally, it is believed that the carbohydrate moieties of HIV-1 gp120 confer the immunosuppressive effects on MoDCs (63). The detection of glycosylation on the minor fimbriae might enable P. gingivalis to prompt its immunosuppressive phenotype in a similar manner.
In conclusion, our results show distinct immunomodulatory roles for two adhesins expressed by the mucosal pathogen P. gingivalis. We show the major fimbriae are immunostimulatory and the minor fimbriae are immunosuppressive. Although co-expressed in wild type strains under laboratory conditions, the two fimbriae are regulated by environmental conditions of direct relevance to their preferred niche. Overall these results may help explain how this oral mucosal pathogen evades and/or suppresses the mucosal immune response, i.e. by uncoupling DC maturation from the cytokine response, leading to anergy.
HIV-1 gp120 CM envelope protein (Cat #2968) was obtained through the NIH AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH:. We would like to thank Dr. Caroline A Genco for providing us with the isogenic fimbriae mutants (DPG-3, MFI, MFB). We would like to thank Dr. D. R. Littman for providing us with the Raji and the Raji DC-SIGN cell lines. We would like to thank Todd Rueb at the Research Flow Cytometry Core facility from Stony Brook University Medical Center, for his input and technical expertise and Baljit Moonga for proofreading the manuscript.
†This study was supported by US Public Health Service grants from the NIH/NIDCR (R01 DE014328 to C.W.C., F31DE020014 to A.Z.)