ABSTRACT
The innate and the adaptive arms of the mucosal immune system must be coordinated to facilitate the control of pathogenic invasion while maintaining immune homeostasis. Toll-like receptors, able to activate the cell to produce bactericidal and inflammatory cytokines but also able to upregulate antigen (Ag)-presenting and costimulatory molecules, are particularly important in this regard. We have previously shown that the chronically infected oral mucosa is in a state of endotoxin tolerance, as evidenced by the downregulation of Toll-like receptors 2 and 4 and of inflammatory cytokines and the upregulation of SH2-containing inositol phosphatase, an inhibitor of NF-κB signaling. In the present study, we hypothesized that endotoxin tolerance would influence the ability of human macrophages to engage in Ag capture and killing of the oral pathogen Porphyromonas gingivalis and to upregulate costimulatory molecules and stimulate autologous T-cell proliferation. We show that uptake, but not killing, of P. gingivalis 381 is enhanced by endotoxin tolerance. Reduced killing is possibly due to a reduction of the intracellular lysosomes. We further show that the expression of the Ag-presenting molecule HLA-DR and costimulatory molecules CD40 and CD86 is dampened by endotoxin tolerance to the constitutive level. This, along with our previous evidence for reduction in immunostimulatory cytokines, is consistent with the observed decrease in the induction of autologous CD4+ T-cell proliferation by endotoxin-tolerized macrophages. Overall, these studies suggest that endotoxin tolerance, as observed in the inflamed oral mucosa, potentiates the innate Ag capture activity of macrophages but diminishes the potential of human macrophages to initiate the adaptive immune response. In conclusion, endotoxin tolerance, while helpful in bacterial clearance and in surmounting excessive inflammatory tissue damage, could potentially reduce the (protective) adaptive immune response during chronic infections such as periodontitis.
Coordination of the innate and the adaptive arms of the immune response is required for the efficient and effective elimination of infectious agents (1, 46). The Toll-like receptors (TLRs) of the innate immune system recognize distinct microbial signatures known as pathogen-associated molecular patterns and marshal an immediate response in order to destroy the invading pathogens, including the induction of inflammatory cytokines and nitric oxide (41) and the release of reactive oxygen species (34). Moreover, TLR-mediated signaling leads to phagosome maturation in antigen (Ag)-presenting cells (APCs) (3) through the formation and fusion of lysosomes to facilitate the efficient loading of processed antigenic peptides on major histocompatibility complex (MHC) class II molecules. Furthermore, the transcription and secretion of immunostimulatory cytokines (29, 40) and of chemokines (45) and the expression of costimulatory molecules such as CD40, CD86, and CD80 (3, 13) are initiated by TLR ligation. These events all help to elicit and direct the adaptive immune response, including the activation of specific subsets of naive T cells (3, 36).
Mucosal surfaces such as the gut and oral mucosa are exposed to an enormous microbial burden (19, 30), one that would overwhelm immune homeostasis if not for various mechanisms in place to regulate and dampen the TLR-mediated inflammatory response (5, 12, 50). Recent studies have identified endotoxin tolerance as one such negative regulatory mechanism that is activated in mucosa of oral (25, 26), gastrointestinal (24, 28), ocular (2, 43), renal (16, 22), and respiratory (15) systems. Endotoxin tolerance blunts the inflammatory response and potentially protects the host from exuberant host tissue damage. The loss of this regulatory mechanism is associated with excessive systemic (4) and mucosal (35) inflammatory disease. Our previous studies suggest that the inflamed oral mucosa is subjected to negative regulation by endotoxin tolerance, as evidenced by the downregulation of TLR2 and TLR4 and the upregulation of negative regulator of intracellular signaling SH2-containing inositol phosphatase in situ. Moreover, we have shown that in vitro endotoxin-tolerized human monocytes/macrophages (MΦ) are impaired in their ability to produce inflammatory cytokines but not anti-inflammatory cytokine interleukin-10 (25, 26).
The present study was designed to determine whether endotoxin tolerance in vitro alters the innate and adaptive immune functions of human MΦ, namely, bacterial uptake and killing, formation of lysosomes, upregulation of costimulatory molecules, and induction of autologous T-cell proliferation. We show that a single lipopolysaccharide (LPS) stimulus enhances the uptake of Porphyromonas gingivalis 381; moreover, when followed by challenge with the same LPS (i.e., induction of homotolerance), uptake is further enhanced. However, killing of P. gingivalis, as determined by use of Sytox+P. gingivalis, was not enhanced by single-LPS stimulus or by endotoxin tolerance, possibly due to a reduction in lysosomal granules. We further show that in endotoxin-tolerized MΦ, the expression of HLA-DR and CD86 is dampened to the level of untreated controls and that the ability of MΦ to stimulate the proliferation of autologous CD4+ T cells is dampened as well, but not to the level of untreated controls.
MATERIALS AND METHODS
Anaerobic bacterial culture and LPS isolation. P. gingivalis wild-type strain 381 was maintained on anaerobic blood agar (Fischer Scientific Co., Springfield, NJ). Cultures were maintained at 37°C in an anaerobic glove box (Coy Laboratory Products, Inc., Ann Arbor, MI) in an atmosphere of 85% N2-5% H2-10% CO2 for 3 to 5 days. Bacteria were cultured until the late log phase of growth. LPS was isolated from either P. gingivalis strain 381 (Pg LPS) or type ATCC strain Escherichia coli 25922 by hot phenol-water extraction followed by isopycnic density gradient centrifugation and was further purified of contaminating nucleic acids, proteins, and lipoproteins. In addition, some of the LPS preparations (purified identically) were a gift (T. E. Van Dyke, Boston University Goldman School of Dental Medicine). The purity of LPS was confirmed by gel electrophoresis, which detected visible LPS ladder staining pattern with silver staining and no visible proteins by Coomassie blue staining (data not shown).
Human peripheral blood monocyte-derived MΦ.Monocytes were isolated from mononuclear fractions of peripheral blood of healthy donors by means of adherence to polystyrene culture flasks as previously described (26). Briefly, whole peripheral blood was centrifuged on Ficoll, and the mononuclear cell fraction pelleted and resuspended in RPMI 1640 (Invitrogen) with 10% heat-inactivated fetal calf serum (Sigma, St. Louis, MO). Mononuclear cells (∼2.7 × 108) were seeded on 150-ml polystyrene culture flasks and incubated for 2 h at 37°C in a 5% CO2 incubator. After the nonadherent cells were washed off, the adherent monocytes were retrieved from the culture dishes by use of a 1× trypsin-EDTA solution. Monocytes were cultured in RPMI 1640 with 10% heat-inactivated fetal calf serum. The percentages of viable monocytes (typically >90% after LPS stimulation) were monitored by trypan blue exclusion. The phenotype of the isolated blood monocytes was established through the expression of CD14 by flow cytometry analysis (data not shown). Human blood-derived monocytes were cultured in the presence of MΦ colony-stimulating factor (M-CSF) for 5 to 7 days. The differentiation of monocytes into MΦ was confirmed by the uptake of 1,1′-dioctadecyl-3,3,3′,3′-tetramethyl-indocarbocyanine perchlorate (Dil)-conjugated labeled low-density lipoproteins (LDLs) by day 7 of culture and by high constitutive expression of HLA-DR by MΦ relative to what was seen for monocytes. The percentages of MΦ (typically >95%) were determined by counterstaining the nuclei with DAPI (4′,6′-diamidino-2-phenylindole) and calculating the percentages of Dil-LDL+ cells over the total DAPI+ cells per ×20-magnification microscopic field.
Autologous T-cell isolation and MΦ cocultures.Autologous CD4+ cells were isolated from the peripheral blood-derived mononuclear component of the same donor that was used for monocyte isolation and MΦ generation. Cells bearing CD4 surface markers were isolated from the mononuclear fraction through negative selection with a cocktail of monoclonal antibodies conjugated with microbeads (Miltenyi Biotec, Gladbach, Germany). The isolation of T cells by negative selection with Mini-macs separation columns (Miltenyi Biotec), as described by the manufacturer, obviated the activation of the CD4+ helper T lymphocytes during the purification procedure. The isolated cells were detected by flow cytometry using fluorescein isothiocyanate-CD4 monoclonal antibody staining. The purity of the CD4+ T cells was more than 99% relative to what was seen for isotype-matched (negative) antibody staining, as analyzed by flow cytometry.
In vitro LPS stimulation and challenge: induction of endotoxin tolerance.Our in vitro design for the induction of endotoxin tolerance has previously been described in detail (25, 26). Briefly, human MΦ were incubated for 24 h either with no LPS stimulus (control) or with initial sensitization with 1,000 ng/ml of P. gingivalis or E. coli LPS for 24 h followed by a challenge with the same LPS (i.e., homotolerance) or with a different LPS (i.e., heterotolerance) at the same initial dosage for a further duration of 24 h. The cells were then pelleted and washed with cold phosphate-buffered saline three times to be used for subsequent experiments as described below.
Standardization of bacterial and MΦ staining with flow cytometry-compatible fluorochromes.The following fluorochromes (Molecular Probes-Invitrogen) were used according to the manufacturer's instructions to label the bacteria for flow cytometry analysis and to quantitate the extent of microbial uptake and killing by MΦ: (i) Syto 64 red fluorescent nucleic acid stain was employed to detect total bacteria associated with MΦ; (ii) Sytox green nucleic acid stain (S-7020) was used to specifically stain the nonviable/killed bacteria associated with MΦ; and (iii) LysoTracker (red DND-99, 577 nm/590 nm [absorbance/emission]) was used to stain acidic lysosomal compartments. Syto-labeled P. gingivalis strain 381 cells at early log phase of growth were added to LPS-treated MΦ at a multiplicity of infection of 25:1 for a period of 3 h. The MΦ were washed three times, and a 50-μl aliquot of the cells in suspension was used for cytocentrifugation (not shown) and subsequent microscopic analysis. The remaining cells were fixed in 1% formaldehyde and stored in the dark until used for flow cytometry. To distinguish adherence from true internalization, trypan blue was added to quench surface fluorescence resulting from extracellular adherent bacteria. Moreover, C-type lectin receptor-mediated uptake, which is dependent on Ca2+, was inhibited with 0.2 mM EDTA to further confirm Ag capture, as reported previously (21). To determine the level of intracellular killing of P. gingivalis by MΦ, a modification of a published protocol (i.e., for bacterial killing by human neutrophils) was used here (7).
Flow cytometry analysis of costimulatory expression on MΦ.Unstimulated MΦ or single P. gingivalis/E. coli LPS-sensitized or -tolerized MΦ were incubated with monoclonal antibodies (or their respective isotype-matched controls) to analyze the regulation of the surface protein (HLA-DR, CD40, CD80, and CD86) expression for a duration of 30 min at 4°C and washed, and the cells were fixed in 1% paraformaldehyde and stored in the dark until fluorescence-activated cell sorting (FACS) analysis. Analysis was performed with FACSCalibur (Becton Dickinson). The expression of the surface proteins was analyzed as the percentages of positive cells in the relevant population as defined by forward-scatter and side-scatter characteristics. Expression levels were evaluated by assessing mean fluorescence intensity indices calculated by relating the mean fluorescence intensity noted with the relevant monoclonal antibody to that with the isotype control monoclonal antibodies for samples labeled in parallel and acquired using the same setting. All antibodies (obtained from BD Biosciences) that were employed were directly conjugated with either fluorescein isothiocyanate or phycoerythrin fluorochromes.
Flow cytometry-based autologous T-cell proliferation assay.Unstimulated control, single-LPS-sensitized, and endotoxin-tolerized MΦ were cocultured with autologous T lymphocytes in culture media supplemented with human AB serum (Sigma Chemical Co., St. Louis, MO) and 10 μg/ml of bromodeoxyuridine (BrdU). T-cell proliferation was quantitated after cell permeabilization using anti-BrdU APC-conjugated secondary antibody along with CD4 staining through flow cytometry according to the manufacturer's (BD Biosciences) instructions. Optimal T-lymphocyte proliferative response was determined by titration of the ratio of (stimulator) MΦ to (responder) T cells. Unstimulated MΦ were used as a negative control and mitogen (phytohemagglutinin)-treated T cells served as a positive control. P. gingivalis LPS-stimulated MΦ (1,000 ng/ml) were cocultured with autologous T cells at stimulator:responder ratios of 1:10 and 1:100. Based on the efficiency of induction of proliferation, the ratio of 1:10 was employed thereafter to determine the effect of endotoxin tolerance on T-cell proliferation (data not shown). The significance of histogram shifts was confirmed by use of the Kolmogorov-Smirnov test at a P value of <0.05 (not shown).
RESULTS
Enhanced Ag capture function of endotoxin-tolerized MΦ.Our previous in vitro studies of endotoxin tolerance (25, 26) used human monocytes, which are weak APCs relative to MΦ and dendritic cells (DCs) (18). Since our goal here was to determine the influence of endotoxin tolerance on immunostimulatory function, M-CSF-induced MΦ were employed. Shown in Fig. 1 is the phenotype of human MΦ demonstrating expression of Dil-LDL (Fig. 1A) and increased expression of HLA-DR after 7 days of culture of monocytes with M-CSF (Fig. 1D). The MΦ were then pulsed with a single LPS stimulus or stimulus/challenge with LPS from P. gingivalis or E. coli, and then MΦ were washed and analyzed for the uptake of Syto-labeled P. gingivalis. Several approaches described in the Fig. 1 legend were used to rule out extracellular adherent bacteria. The results of FACS analysis (Fig. 2) show that a single stimulation with either LPS enhanced the Ag capture function of MΦ. LPS challenge with the same LPS (i.e., homotolerance) or with a different LPS (heterotolerance) further enhanced the Ag capture ability of MΦ; however, there was a consistent trend for a more potent influence on Ag capture function when MΦ were sensitized first with E. coli LPS compared to initial sensitization with P. gingivalis LPS (Fig. 2E versus 2F), although this was not tested statistically.
Human M-CSF-induced MΦ and standardization of fluorochrome- and FACS-based assays. Human blood monocytes cultured in the presence of M-CSF for 5 to 7 days differentiated into MΦ, as determined by the percentage of LDL-Dil+ (red-stained) cells (A) relative to total DAPI+ cells (B) by day 7 of culture and by high constitutive expression of HLA-DR by MΦ (D) relative to monocytes (C) according to FACS analysis. Standardization of Ag capture, killing assays, and assays of lysosome formation by MΦ were done by a combination of fluorescent microscopy (G to J) and FACS (C to F) analysis. Confirmation of the uptake/internalization of P. gingivalis by MΦ was obtained by several approaches as follows: (i) the use of 0.2 mM EDTA to chelate Ca2+, required for C-type lectin receptor-mediated (non-opsonin-dependent) uptake, which completely eliminated the uptake of P. gingivalis by FACS analysis (E and F); (ii) The addition of trypan blue to quench extracellular bacteria during FACS analysis, which did not quench bacterial uptake (not shown); and (iii) the colocalization of P. gingivalis into lysosomes. DAPI+ MΦ (G) are shown containing lysosomes (red) (H) and Sytox-positive (green) P. gingivalis (I) after 3 h. Lysosomal routing of P. gingivalis is evidenced by the colocalization of P. gingivalis (yellow arrows) with lysosomes (J). The percentages of viable MΦ before and during the experiments (typically >95%) were monitored by trypan blue exclusion.
Ag capture function of MΦ for P. gingivalis enhanced by endotoxin tolerance. In vitro monocyte-derived MΦ were sensitized with a single LPS stimulus or were sensitized and then challenged with LPS from P. gingivalis or E. coli. As described in Materials and Methods, MΦ were analyzed for uptake of Syto-labeled P. gingivalis. Increased uptake is indicated by a shift of the histogram to the right. A single stimulation of MΦ with either LPS enhanced the Ag capture function of P. gingivalis relative to what was seen for the no-LPS control. Furthermore, MΦ subjected to homotolerance (C and D) or heterotolerance (E and F) were further enhanced in Ag capture ability of P. gingivalis. The assay was repeated three times; results are representative of three separate analyses. Ec LPS, sensitization with E. coli LPS; Pg LPS, sensitization with P. gingivalis LPS; EcEc LPS, sensitization and then challenge with E. coli LPS (homotolerance); PgPg LPS, sensitization and then challenge with P. gingivalis LPS (homotolerance); EcPg LPS, sensitization with E. coli LPS and then challenge with P. gingivalis LPS (heterotolerance); PgEc LPS, sensitization with P. gingivalis LPS and then challenge with E. coli LPS (heterotolerance).
Reduction in microbial killing and of lysosomes in MΦ by endotoxin tolerance.Preliminary epifluorescence microscopy and imaging analysis studies established that P. gingivalis was internalized and compartmentalized within lysosomes (Fig. 1H) and became positive for Sytox, suggesting loss of viability in MΦ (Fig. 1J). Here we show that endotoxin tolerance induces a decrease in Sytox+P. gingivalis within MΦ, as determined by FACS analysis (Fig. 3A). Moreover, there is a reduction of intracellular lysosomes in tolerized MΦ, as determined by FACS analysis of LysoTracker+ MΦ (Fig. 3B).
Dampening of microbial killing and lysosomal formation in MΦ induced by endotoxin tolerance. (A) MΦ were subjected to a single stimulus with LPS or, to induce a state of endotoxin tolerance, LPS stimulus followed by LPS challenge as indicated. MΦ from all these groups were then pulsed with P. gingivalis strain 381 and Sytox labeling followed by FACS analysis. A shift to the left or right indicates a decrease or increase, respectively, in Sytox labeling. The induction of endotoxin homotolerance (A1 and A2) or heterotolerance (A3 and A4) resulted in a slight decrease in Sytox+P. gingivalis within MΦ relative to that of the single-LPS sensitization, as determined by flow cytometry analysis. (B) The formation of acidic lysosomes in MΦ was followed by the incorporation of LysoTracker. The results suggest a reduction of intracellular lysosomes in endotoxin-tolerized MΦ relative to what was seen for single-LPS sensitization. The assay was repeated a minimum of two times; results are representative of three separate analyses. See Fig. 2 legend for explanation of abbreviations.
Reduction in MHC class II and in costimulatory molecules on endotoxin-tolerized MΦ.Optimum Ag presentation requires the expression of MHC class II molecules and of costimulatory molecules on APCs (6). We show here that MΦ constitutively express HLA-DR (Fig. 1D and 4A); moreover, HLA-DR expression is upregulated by a single stimulus with LPS of P. gingivalis or E. coli (Fig. 4A). However, endotoxin tolerance induces the downregulation of HLA-DR, to the level of the control in the case of P. gingivalis LPS or below the constitutive level in the case of E. coli LPS. Moreover, MΦ express CD40 constitutively (Fig. 4B) and very low levels of CD86 (Fig. 4C), and CD80 was not expressed at all (not shown). A single stimulus with LPS from P. gingivalis or E. coli resulted in the upregulation of CD40 and CD86; however, the induction of endotoxin tolerance (i.e., challenge with E. coli LPS, but not P. gingivalis LPS) induced the downregulation of CD40 and CD86 (Fig. 4). There was no apparent difference in the influences of homotolerance and heterotolerance on the downregulation of HLA-DR, CD40, and CD86, but an initial stimulation with E. coli LPS was required for optimal dampening of CD40 and CD86 on MΦ.
Decreased expression of MHC class II and CD40 and CD86 on endotoxin-tolerized MΦ. While MΦ constitutively express HLA-DR (no LPS) (A1 and A2), CD40 (B1 and B2), and CD86 (C1 and C2) but not CD80 (not shown), the expression of HLA-DR is upregulated by a single stimulus with LPS of P. gingivalis or E. coli (A1 and A2). However, endotoxin tolerance induces the downregulation of HLA-DR, to the constitutive level (i.e., no-LPS control) in the case of P. gingivalis LPS (A3 and A5) or below the constitutive level as induced by E. coli LPS (A4 and A5). Both CD40 (B) and CD86 (C) were upregulated by a single stimulus with LPS from P. gingivalis (B1 and C1) or E. coli (B2 and C2). The induction of endotoxin tolerance by LPS from E. coli or P. gingivalis induced the downregulation of CD40 (B4 and B6) and CD86 (C4 and C6) but only when stimulated initially with E. coli LPS. Stimulation with LPS from P. gingivalis, followed by challenge with P. gingivalis or E. coli LPS, resulted in no change in the regulation of CD40 (B3 and B5) and slight downregulation of CD86 (C3 and C5). The assay was repeated a minimum of four separate times; typical results are shown. MFI, mean fluorescence intensity; see Fig. 2 legend for explanation of other abbreviations.
Autologous T-cell proliferation induced by MΦ is dampened by endotoxin tolerance.To avoid the influence of MHC class II mismatch on the T-cell response to MΦ, autologous T cells were used as responder cells. This is reflected in the extremely low baseline proliferation of T cells cocultured with unstimulated MΦ (Fig. 5A and B). When MΦ were sensitized with LPS of P. gingivalis or E. coli, T cells were induced to proliferate, as indicated by the uptake of BrdU-APC. However, endotoxin-tolerized MΦ, either homotolerized or heterotolerized, were less efficient at stimulating T cells to proliferate, as evidenced by the reduction in BrdU-APC, compared to single-LPS-stimulated MΦ (Fig. 5C to F).
MΦ-mediated autologous T-cell proliferation dampened by endotoxin tolerance. (A and B) Shown are the extremely low proliferative responses of T cells to unstimulated MΦ. When MΦ were subjected to a single stimulation with LPS of P. gingivalis or E. coli, T cells were induced to proliferate, as indicated by uptake of BrdU-APC. (C to F) Induction of endotoxin tolerance in MΦ (either homotolerance or heterotolerance) resulted in a dampening of T-cell proliferation, indicated by a leftward shift of the histogram. The T-cell proliferation assay was repeated two times; representative results are shown. The significance of histogram shifts was confirmed by use of the Kolmogorov-Smirnov test at a P value of <0.05 (not shown). See Fig. 2 legend for explanation of abbreviations.
DISCUSSION
We show that the induction of endotoxin tolerance in MΦ through LPS stimulus/challenge results in an enhanced capability of MΦ to internalize P. gingivalis (Fig. 2). That this is true internalization and not adherence was confirmed by several approaches, including (i) image-enhanced epifluorescence microscopy, i.e., by colocalization of Sytox+P. gingivalis with lysosomal compartments; (ii) the use of trypan blue, in combination with FACS analysis, to quench extracellular fluorescence; and (iii) chelation of Ca2+, required for nonopsonic bacterial uptake by C-type lectin receptors (21) (Fig. 1). Microbial recognition by MΦ and other APCs is mediated by signaling types of pattern recognition receptors such as TLRs and nucleotide-binding oligomerization domains, while clearance of many bacterial species (possibly including P. gingivalis) is likely mediated by alternative nonsignaling receptors, such as select members of the C-type lectin receptor and scavenger receptor families (9, 31). The C-type lectin receptors are suspected due to evidence for their role in the uptake of P. gingivalis by DCs (unpublished data from our laboratory); however, previous studies indicate that several are negatively regulated on monocytes/MΦ by LPS treatment, including DC-specific ICAM-3-grabbing nonintegrin (33) and MΦ mannose receptor (CD206) (17). In contrast, macaque monocyte-derived DCs stimulated with LPS upregulate the putative Ag uptake receptor DEC-205 (23). Scavenger receptors may also be involved in LPS-enhanced uptake. TLR ligands specifically promote bacterial phagocytosis in both murine and human cells through induction of the phagocytic gene program. This involves induction of MyD88 (myeloid differentiation factor 88)-dependent signaling through interleukin-1 receptor-associated kinase-4 and p38, leading to the upregulation of scavenger receptors MARCO (MΦ SR with collagenous structure), LOX-1 (lectin-like oxidized LDL receptor), and SR-A (scavenger receptor-A) (8). LPS induces a marked increase in SR-A (CD163) expression on circulating monocytes 24 h following experimental endotoxemia. CD163 is mostly known for its uptake of hemoglobin/haptoglobin complexes, and its expression is increased by interleukin-10 (14), an anti-inflammatory cytokine released by endotoxin-tolerized monocytes (26). SR-A is involved in the direct binding and uptake of various gram-positive and gram-negative organisms (31); however, previous studies suggested that cultured human monocyte-derived MΦ downregulate SR-A expression in vitro when exposed to LPS (10). Interestingly, the same report indicated that primary and elicited mouse peritoneal MΦ as well as J774A.1 and RAW264.7 mouse MΦ lines increase SR-A expression in response to LPS. Further study is required to identify the Ag capture receptor(s) upregulated by endotoxin tolerance.
Our data further suggest that despite enhanced Ag capture induced by endotoxin tolerance, the overall killing ability of MΦ, as evidenced by internalized Sytox+P. gingivalis, is decreased (Fig. 3A). This correlates with a reduction in detectable lysosomes, revealed by FACS analysis of lysosomotropic dye (LysoTracker; Molecular Probes) in MΦ that have been endotoxin tolerized with either P. gingivalis or E. coli LPS (Fig. 3B). Earlier studies in rabbits have shown that single exposure to endotoxin results in extracellular liberation of lysosomal enzymes (39), but the endotoxin tolerance model was not employed. Nor was the effect of endotoxin tolerance analyzed in several more recent conflicting studies that suggest that TLR signaling may speed or slow phagosome maturation (reviewed in reference 44). Using MΦ from mice lacking TLR2, TLR4, and the TLR signaling adaptor molecule MyD88, Blander and Medzhitov explored the fate of phagosomes containing gram-negative or gram-positive bacteria (3). Acquisition of lysosomal markers (LAMP-2 protein and LysoTracker) was significantly impaired in the absence of TLR signaling. The authors suggested that TLR signaling was necessary to permit phagosome/lysosome fusion. During the phagocytosis of apoptotic cells, however, phagosomes acquired lysosomal markers at equivalent rates in wild-type and TLR-deficient MΦ, and TLR stimulation of wild-type cells did not enhance the rate of maturation of the phagosomes, suggesting that TLR signaling speeds the rate of maturation of the phagosome compartment. In a contrasting study (36), two lysosomal tracers were employed to measure phagosome/lysosome fusion in mouse MΦ to determine the internalization of a variety of particles. The latter study reported that the rate of phagosome/lysosome fusion was increased in MΦ lacking TLR4. In this case, TLR4 appeared to lower the rate of phagosome maturation, independent of the presence of any obvious TLR4 ligand. The mechanism by which TLR4 might regulate phagosome maturation in the absence of a ligand is not clear. Phagosomal pH is associated intimately with phagosome maturation (11, 38). The inhibition of acidification reduces phagosome maturation and the acquisition of lysosomal constituents; conversely, the inhibition of phagosome maturation reduces its acidification and, generally, its ability to process internalized Ags (37, 38, 49). Certain LPS moieties (e.g., LPS O side chain of Brucella spp.) are involved in inhibition of the early fusion between Brucella suis-containing phagosomes and lysosomes in murine MΦ (32). Thus, there is an extensive body of data implying that TLR signaling impacts on many facets of particle binding, uptake, and degradation; however, a dearth of data exists on endotoxin tolerance and Ag capture and killing.
It is interesting that P. gingivalis LPS-initiated heterotolerance, i.e., P. gingivalis LPS sensitized, E. coli LPS challenged, did not mediate as potent an effect on Ag capture, lysosomes, or maturation as did E. coli LPS-initiated heterotolerance (E. coli LPS sensitized, P. gingivalis LPS challenged). One possible explanation for this difference is due to the relatively narrower range of signaling mediated through TLR2, the receptor involved in recognition of P. gingivalis LPS. E. coli LPS, which is recognized through TLR4, has a wider armamentarium mediated through all the four known adaptor molecules (12), which eventually recruited the downstream kinases and transcription factors.
Our results further show that the surface expression of Ag-presenting molecule HLA-DR, as well as CD40 and CD86, on human MΦ is downregulated by endotoxin tolerance (Fig. 4). This is likely due to the inhibition of p38 mitogen-activated protein kinase by endotoxin tolerance. The inhibition of P38 mitogen-activated protein kinase has been previously described to downregulate HLA-DR and costimulatory molecules and to reduce the allostimulatory activity of DC (27). Moreover, the ability of endotoxin-tolerized MΦ to stimulate autologous CD4+ T cells to proliferate is dampened (Fig. 5). The role of septic shock in inducing dramatic downregulation of MHC class II surface expression on monocytes and an impairment of Ag presentation capacity and of Ag-specific immunity (47, 48) has been described previously. Endotoxin tolerance is involved in blunting (harmful) inflammatory responses at mucosal surfaces of oral (25, 26), gastrointestinal (24, 28), ocular (2, 43), renal (16, 22), and respiratory (15) systems. These mucosal systems are constantly exposed to a high burden of commensal and pathogenic microbes (19, 30). Commensal species appear to prevent the colonization of pathogens and to stimulate the development of mucosally associated lymphoid tissue (20, 42).
In summary, endotoxin tolerance represents the proverbial “double-edged sword,” enhancing the ability of APCs to engage in Ag capture and preserving mucosal immune homeostasis but impairing the activities involved in Ag processing and presentation. Based on previous evidence for the induction of endotoxin tolerance in inflamed oral mucosa (25, 26), it is still unclear whether endotoxin tolerance is predominantly beneficial, or whether it prevents the development of a strong adaptive immune response in chronic periodontitis.
ACKNOWLEDGMENTS
This study was supported by a U.S. Public Health Service grant from the NIH/NIDCR (R01 DE14328).
FOOTNOTES
- Received 18 January 2007.
- Returned for modification 14 March 2007.
- Accepted 29 October 2007.
- Copyright © 2008 American Society for Microbiology