INTRODUCTION

Periodontitis is a chronic inflammatory disease caused by bacteria, which is one of the two major oral diseases in humans. It is accompanied by the destruction of tooth-supporting structures and is difficult to treat [1]. Porphyromonas gingivalis (P. gingivalis), a gram-negative bacteria with black pigmentation, has been regarded as one of the most important periodontopathic bacteria [2]. As a major constituent of the outer membrane of P. gingivalis, lipopolysaccharide (LPS) is considered to be a potential inducer of proinflammatory cytokines, anti-inflammatory cytokines, and chemokines [3]. Appropriate host responses to P. gingivalis LPS are essential to resist invading bacteria, but excessive and prolonged immune responses can lead to the destruction of periodontal tissues [4].

Endotoxin tolerance is defined by a reduced capacity of the host to respond to subsequent LPS activation following an exposure to initial stimulus and is characterized by diminished release of proinflammatory cytokines, such as TNF-α and IL-1ß [5]. Hyporesponsiveness to a secondary challenge with the same LPS is referred to as homotolerance, while that with a different LPS is termed heterotolerance [6]. These two kinds of tolerances serve as important pathophysiological adaptations to regulate over-exuberant inflammation against bacteria, but they may also favor further infections [7]. Molecular mechanisms of endotoxin tolerance, especially for those in periodontitis, still need to be fully elucidated.

Toll-like receptors (TLRs) is a family of mammalian homologs of Drosophlia Toll [8]. Among them, TLR2 and 4 function as the principal innate sensors for virulence factors of periodontopathic bacteria and might be very important in endotoxin tolerance developed in periodontal tissues [9, 10].

TLR2 and 4 signaling pathways involve a cascade of adaptors, intermediates, and effectors, including myeloid differentiation factor-88 (MyD88), IL-1 receptor-associated kinases (IRAKs), and TNF receptor-associated factor-6 (TRAF6) [11]. The mammalian family of IRAK molecules contains four members (IRAK1, IRAK2, IRAK-M, and IRAK4). Among them, IRAK4 is essential for the phosphorylation of IRAK1. IRAK-M may act as a negative regulator by preventing the dissociation of IRAK1 and IRAK4 from MyD88 and the subsequent formation of its complex with TRAF6 [12].

Toll-interacting protein (Tollip) is identified initially as an inhibitory signaling molecule in interleukin-1 signaling. It functions through its ability to suppress potently the activity of IRAK after the activation of TLRs and therefore serve to limit the production of proinflammatory cytokines [13]. Suppressor of cytokine-signaling-1 (SOCS1) is another recently reported negative regulator in TLR2 and 4 signaling pathways, which is related with the inactivation of NF-κB [14].

In this present study, we hypothesized that TLR2, TLR4 and some of the signaling intermediates, including IRAK4, IRAK-M, Tollip and SOCS1, might be involved in endotoxin tolerance induced by P. gingivalis LPS and (or) Escherichia coli (E. coli) LPS, the most classic gram-negative bacterial LPS.

MATERIALS AND METHODS

Reagents

P. gingivalis ATCC 33277 LPS and E. coli O127:B8 LPS were purchased from InvivoGen (CA, USA) and Sigma Aldrich (MI, USA) separately. ELISA kits were provided by BioSource (CO, USA). Fluorescein isothiocyanate (FITC)-conjugated anti-TLR2 antibody and PE-conjugated anti-TLR4 antibody were supplied by eBioscience (CA, USA). Antibodies to IRAK4 and IRAK-M were obtained from Cell Signaling (MA, USA).

Cell Culture and Tolerance Induction

THP-1 cells were cultured in 6-well plates (5 × 105 cells/ml) and divided into seven groups (n = 5 per group). Group 1 was incubated in medium alone. Groups 2 and 5 were cultured in medium for 24 h, washed three times with fresh medium, and stimulated with 1 μg/ml P. gingivalis LPS or 1 μg/ml E. coli LPS, respectively, for 4 h to perform western blot, 6 h for real-time PCR, and 24 h for ELISA and flow cytometry. Groups 3, 4, 6, and 7 were pretreated with 1 μg/ml P. gingivalis LPS or 1 μg/ml E. coli LPS for 24 h, washed, and challenged again with either 1 μg/ml P. gingivalis LPS or 1 μg/ml E.coli LPS for the subsequent detections as described in groups 2 and 5.

Cytokine Detection

At the end of the culture period, supernatants were harvested and stored at −20 °C. Levels of TNF-α, IL-1β, IL-8, and IL-10 in the culture supernatants were determined by commercial ELISA kits according to the manufacturer's instructions.

Flow Cytometry

THP-1 cells were collected and incubated with FITC-conjugated anti-TLR2 antibody and phycoerythrin (PE)-conjugated anti-TLR4 antibody for 30 min at 4 °C in the dark. FITC-conjugated IgG2b and PE-conjugated IgG1 were served as corresponding isotype controls for nonspecific binding of antibodies. After this incubation, cells were washed and then fixed in 1 % paraformaldehyde [15]. Expressions of TLR2 and 4 on 10,000 viable cells were gated and analyzed by a FACSCalibur (BD Biosciences, USA).

Western Blot

Cells were harvested in ice-cold lysis buffer consisting of 1 % Triton X-100, 20-mM Tris and inhibitors for proteinase. Protein extracts were then separated by electrophoresis in 10 % SDS-polyacrylamide gels, and transferred to a PVDF membrane. The membrane was blocked with PBS containing 0.1 % Tween-20 and 5 % nonfat milk, and incubated with antibodies against IRAK4, IRAK-M, and ß-actin overnight. After incubation with appropriate horseradish peroxidase-conjugated secondary antibodies for 1 h, immunoreactive proteins were visualized using a chemiluminescent ECL assay kit (Thermo, USA) [16].

Real-Time PCR

Total RNA was prepared from THP-1 cells stimulated with or without LPS. cDNA was synthesized using a reverse transcription kit (Takara, China). Levels of ß-actin mRNA served as internal controls. The primer sequences were as follows (F/R): Tollip (CAGGCGTAGACTCTTTCTATCTC/GACTCCGAGATGGTGATGTG); SOCS1 (AACTGCTCTTTCGCCCTTAGC/CAGCTCGGAGAGGCAGTCG); and ß-actin (CCTGTACGCCAACACAGTGC/ATACTCCTGCTTGCTGATCC).

Real-time PCR analysis was performed in duplicates in an ABI PRISM 7300 Real-Time PCR System (Applied Biosystems, USA) with SYBR Premix EX Taq (Takara, China). The reaction product was quantified by the standard curve method [3].

Statistical Analysis

Statistical analysis of ELISA data was performed using ANOVA. The data of flow cytometry and real-time PCR were analyzed using the Kruskal–Wallis test, and subsequently, the Mann–Whitney test was performed as a post hoc test. The level of significance was set at p < 0.05.

RESULTS

Cytokine Production in THP-1 Cells upon a Primary or Secondary Exposure to LPS

Stimulations with P. gingivalis LPS or E. coli LPS for 24 h resulted in an enhanced secretion of TNF-α, IL-1β, IL-10, and IL-8 (p < 0.05), and the amounts of all cytokines induced by P. gingivalis were significantly lower than those induced by E. coli LPS (p < 0.05) (Fig. 1).

Fig. 1
figure 1

Cytokine production in THP-1 cells stimulated with LPS. THP-1 cells were pretreated with medium, 1 μg/ml P. gingivalis LPS or 1 μg/ml E. coli LPS for 24 h, washed, and then incubated with medium, 1 μg/ml P. gingivalis LPS or 1 μg/ml E. coli LPS for another 24 h. The levels of TNF-α (a), IL-1β (b), IL-8 (c), and IL-10 (d) in the culture supernatants were measured by ELISA. Data are expressed as mean ± SD (n = 5 per group). *p < 0.05.

Full size image

After restimulation with the same LPS for additional 24 h, the levels of TNF-α and IL-1β were decreased markedly compared with those from THP-1 cells stimulated with P. gingivalis LPS or E. coli LPS only once (p < 0.05), while the production of IL-10 was increased significantly (p < 0.05). Interestingly, a serve reduction in the levels of IL-8 was only observed in cells challenged with repeated E. coli LPS but not P. gingivalis LPS, which demonstrated the differences in endotoxin tolerance drove by P. gingivalis LPS and E. coli LPS (p < 0.05) (Fig. 1).

In heterotolerance groups (groups 4 and 7), a noticeable downregulation of TNF-α and IL-1β was also detected after retreatment with different LPS (p < 0.05), while there were no significant differences at the levels of IL-10 and IL-8 (p > 0.05). Moreover, the amounts of TNF-α and IL-1β in homotolerance groups were significantly lower than those in heterotolerance groups (p < 0.05), except IL-1β production between groups 3 and 4 (p > 0.05) (Fig. 1).

Expressions of TLR2 and 4 in THP-1 Cells after a Primary or Secondary LPS Exposure

Upregulation of TLR2 or TLR4 protein was observed in cells stimulated with P. gingivalis LPS or E.coli LPS, respectively (p < 0.05), which demonstrated the different abilities of these two kinds of LPS to modulate expression levels of members of the TLR family (Fig. 2).

Fig. 2
figure 2

Protein expression changes of TLR2 and 4 in THP-1 cells stimulated with LPS. THP-1 cells were stimulated with medium or LPS as described in the legends to Fig. 1. Protein expression levels of TLR2 (FITC-conjugated) (a) and TLR4 (PE-conjugated) (b) were detected by flow cytometry. Data are expressed as mean ± SD (n = 5 per group). *p < 0.05. One representative result of five independent experiments is shown (c).

Full size image

Furthermore, a significant downregulation of TLR2 protein was detected in THP-1 cells pretreated with either kind of LPS and challenged with P. gingivalis LPS again (p < 0.05). A marked decrease in TLR4 expression was also observed in cells preconditioned with any kind of LPS and retreated with E. coli LPS (p < 0.05). At the levels of both TLR2 and TLR4, there were no significant differences between homotolerance groups and heterotolerance groups (p > 0.05) (Fig. 2).

Differential Expressions of IRAK-M but not IRAK4 in Endotoxin Tolerance

After the first P. gingivalis LPS or E. coli LPS stimulation, both IRAK4 and IRAK-M expressions were increased compared with those without LPS challenge. In addition, after P. gingivalis LPS or E. coli LPS retreatment, the expression levels of IRAK-M were increased in both homotolerance groups and heterotolerance groups, but no expression changes could be observed at the levels of IRAK4 (Fig. 3).

Fig. 3
figure 3

Protein expressions of IRAK4 and IRAK-M in THP-1 cells stimulated with LPS detected by western blot. THP-1 cells were pretreated with medium, 1 μg/ml P. gingivalis LPS or 1 μg/ml E. coli LPS for 24 h, washed, and then challenged with medium, 1 μg/ml P. gingivalis LPS or 1 μg/ml E. coli LPS for another 4 h. Western blot was used to explore IRAK4 (a) and IRAK-M (b) protein expressions. Levels of ß-actin served as internal controls. One representative result of three independent experiments is shown.

Full size image

Enhanced Expressions of SOCS1 but not Tollip Contributed to the Induction of Endotoxin Tolerance

After the first 24-h challenge, both kinds of LPS led to the significant increases in Tollip and SOCS1 mRNA expressions (p < 0.05). The levels of Tollip induced by P. gingivalis LPS were significantly higher than those induced by E. coli LPS (p < 0.05), while the differences in SOCS1 expression levels between these two groups weren't statistically significant (p > 0.05) (Fig. 4).

Fig. 4
figure 4

Gene expression changes of Tollip and SOCS1 in THP-1 cells stimulated with LPS. THP-1 cells were pretreated with medium, 1 μg/ml P. gingivalis LPS or 1 μg/ml E. coli LPS for 24 h, washed, and then restimulated with medium, 1 μg/ml P. gingivalis LPS or 1 μg/ml E. coli LPS for 6 h. Real-time PCR was used to quantify Tollip (a) and SOCS1 (b) mRNA expression levels. The absolute mRNA levels of all the genes were normalized to ß-actin levels of individual samples. Data are expressed as mean ± SD (n = 5 per group). *p < 0.05.

Full size image

After pretreatment with P. gingivalis LPS or E. coli LPS for 24 h and a subsequent restimulation for additional 6 h, the expressions of SOCS1 in both homotolerance groups and heterotolerance groups were significantly higher than those in cells stimulated only once (p < 0.05), but no significant differences were observed between these two groups (p > 0.05). In contrast, the expression levels of Tollip were only increased in THP-1 cells preconditioned with E. coli LPS and subsequently stimulated with P. gingivalis LPS (p < 0.05) (Fig. 4).

DISCUSSION

In this present study, we evaluated the cytokine production in THP-1 cells upon repeated P. gingivalis LPS and (or) E.coli LPS challenges, indicating the differences in endotoxin tolerance triggered by these two kinds of LPS at the levels of several cytokines. In addition, we disclosed the involvement of TLR2 and 4 signaling pathways, especially the negative regulators, IRAK-M and SOCS1, in P. gingivalis LPS tolerization for the first time.

It is generally accepted that the primary etiologic factor of periodontitis is bacterial biofilm and specific microorganisms in subgingival plaque, including P. gingivalis, Prevotella intermedia (P. intermedia), Fusobacterium nucleatum (F. nucleatum), and Aggregatibacter actinomycetemcomitans (A. actinomycetemcomitans), initiate the disease [17]. LPS is one of the most important virulence factors of gram-negative bacteria, and could cause direct destruction of periodontal tissues and stimulate the host to trigger a wide range of inflammatory responses [18, 19]. In comparison to the classic gram-negative bacterial LPS, E. coli LPS, P. gingivalis LPS is unique in its biological activities and pathogenicities. Its lipid A is a monophosphate type lacking a phosphate group in the 4′ position that does not consist of tetradecanoic acids, but rather long-chain fatty acids made up of only acyloxyacyl groups. Therefore, endotoxic activities of P. gingivalis LPS is relatively weaker, and some researchers believed that P. gingivalis LPS could act on host cells through TLR2, but not TLR4 just as what was found in E. coli LPS [10, 20]. Furthermore, accumulated evidences suggested that LPS from some other periodontopathic bacteria, such as P. intermedia, F. nucleatum, and A. actinomycetemcomitans, could activate TLR4, and there were some similarities in potential for triggering inflammation between E. coli LPS and these periodontopathic bacterial LPS [21, 22]. Therefore, in this experiment, we chose two different kinds of LPS, P. gingivalis LPS and E. coli LPS, as the stimulators to explore the host responses to endotoxin tolerance, which might take place in the development of periodontal inflammation.

TNF-α is probably the best marker for endotoxin tolerance [23] and many researchers reported the reduced secretion of TNF-α in the cells upon repeated LPS challenge [24, 25]. However, endotoxin tolerance doesn't mean the global decline of all cytokines. Instead, it represents a selective reprogramming aimed at limiting inflammatory damage [26]. The reasons for differential cytokine expressions might be very complicated. We speculate it might be associated with different signaling pathways for different cytokine production.

Present results from our group also showed that the levels of TNF-α and IL-1β in the cells from homotolerance groups were much lower than those from heterotolerance groups. As for the production of IL-10, there was a significant change in homotolerance groups rather than heterotolerance groups. Therefore, hyporesponsiveness to a different subsequent stimulus might be usually weaker than that to the same LPS.

According to the important roles of TLR2 and 4 in recognizing pathogen associated molecular patterns, we paid attention to their possible involvement in tolerance. Our results revealed that in response to repeated P. gingivalis LPS or E. coli LPS challenge, TLR2 or 4 expressions were down-regulated, which might inhibit excessive signaling transduction activated by persistent bacteria stimulations. Distinct expressions of TLR2 and 4 suggested the different regulatory mechanisms, which might be responsible for the different expression profiles of cytokines secreted by P. gingivalis LPS- or E. coli LPS-tolerized cells. E. coli LPS, as a TLR4 agonist, activates both MyD88-dependent and MyD88-independent pathways and down-regulates cytokine production in endotoxin-tolerant cells, while TLR2 signals only through MyD88-dependent pathway. Moreover, impaired and partial endotoxin tolerance induced in monocytes/macrophages by P. gingivalis LPS, characterized by persistent high secretion of IL-8 might be due to the lack of immunomodulatory effect of IFN-β [27].

Similar to our results, Li et al.'s research also confirmed that pretreatment of THP-1 cells with bacterial lipoprotein strongly inhibited TLR2 activation in response to subsequent stimulation [28]. Interestingly, Savidge et al. demonstrated that chronic exposure of intestinal epithelial cells (IEC) to LPS-induced tolerance at the levels of IL-8 independently of TLR4 signaling [29]. A possible explanation for this discrepancy might be the variation in TLR4 expression levels between THP-1 cells and IEC.

In this present research, we only explored protein expression levels of TLR2 and 4, but not activation states of these two receptors. Chen believed that LPS-induced NF-κB activation and cytokine gene expressions required tyrosine phosphorylation of TLR4 [30]. Tyrosine phosphorylation can regulate signal transducing functions of many receptors, including TLR2, TLR3, and TLR4, which leads to post-translational modifications within the TIR domain and plays an important role in the recruiting of MyD88 and IRAKs [31, 32]. Therefore, further studies are needed to explore tyrosine phosphorylation of TLR2 and 4 in endotoxin-tolerant cells.

Mechanisms of heterotolerance might be more complex than those of homotolerance. Actually, periodontal infections are polymicrobial, and numerous virulence factors are involved in it. Endotoxin tolerance in periodontal tissues is a kind of heterotolerance rather than homotolerance. We presumed that there might be crosstalk between TLR2 and TLR4 pathways and some key signaling intermediates in the downstream of TLR2 and 4 might play a role in endotoxin tolerance, especially in heterotolerance.

Function of IRAKs is also highly regulated via multiple mechanisms, such as the positive regulator IRAK4 and the negative regulator IRAK-M [12]. In our present research, no significant differences were confirmed in IRAK4 protein expressions before and after P. gingivalis LPS retreatment. Xiong et al. believed that induction of endotoxin tolerance didn't suppress the total expression levels of IRAK4 in THP-1 cells and human monocytes. Instead, tolerance impaired recruitment of IRAK4 to TLR4 and activation of IRAK4 [33]. In contrast to the comparable expressions of IRAK4, increased levels of IRAK-M in our experiment disclosed their contribution to the development of endotoxin tolerance. After LPS challenge, the complex formed by MyD88, IRAK1, IRAK4, and TRAF6 induces a series of phosphorylation events that lead to the activation of NF-κB and activator protein 1. It has been suggested that IRAK-M could block the signaling between MyD88 and IRAK1 by binding to the complex, and then inhibit the resulting cytokine production [34].

Tollip is an endogenously expressed molecule that limits proinflammatory signals from both innate and adaptive immune systems. It might potently suppress the activity of IRAK after the activation of TLRs, either by blocking IRAK autophosphorylation or by interfering with the binding of a crucial signaling molecule, such as a second kinase [35]. To our disappointment, we failed to confirm a distinct upregulation of Tollip mRNA expressions upon repeated P. gingivalis LPS or E. coli LPS treatment. Interestingly, there was an enhanced expression of Tollip in the cells pretreated with E. coli LPS and treated with P. gingivalis LPS. Advanced exploration is needed to disclose whether there is a synergetic effect or crosstalk between these two kinds of LPS signal transduction pathways.

SOCS1 was described as a suppressor factor involved in a coordinated negative regulation of immune responses [36]. Our study revealed that an increase in SCOC1 expressions was related with not only homotolerance but also heterotolerance. SOCS1 could regulate cytokine signaling pathways via inhibition of tyrosine phosphorylation on cytokine receptors and signaling molecules [37]. It also could inhibit NF-κB activation by IRAKs and TRAF6 but not by I-κB kinase-β (IKK-β), indicating that the inhibition by SOCS1 occurred downstream of TRAF6 and upstream of IKK-β [38]. In addition, SOCS1 could interact with Toll/IL-1 receptor-associated protein (TIRAP), which specifically mediates MyD88-dependent pathway via TLR2 and 4, and lead to its degradation. Degradation of signaling proteins, such as TIRAP, severely dampens LPS signaling and impairs cytokine production [39].

In summary, this present study demonstrated the quantitatively and qualitatively different cytokine production in P. gingivalis LPS- or E. coli LPS-tolerized THP-1 cells, and the involvement of TLR2, TLR4, IRAK-M, and SOCS1 in endotoxin tolerance. In addition, this pattern of immune response suggests that the above-mentioned receptors and inhibitory signaling molecules might be potential targets for controlling inflammation and maintaining homeostasis in periodontal tissues.