Porphyromonas gingivalis Stimulates TLR2-PI3K Signaling to Escape Immune Clearance and Induce Bone Resorption Independently of MyD88

Abstract

Porphyromonas gingivalis is a gram-negative anaerobic periodontal pathogen that persists in dysbiotic mixed-species biofilms alongside a dense inflammatory infiltrate of neutrophils and other leukocytes in the subgingival areas of the periodontium. Toll-like receptor 2 (TLR2) mediates the inflammatory response to P. gingivalis and TLR2-deficient mice resist alveolar bone resorption following oral challenge with this organism. Although, MyD88 is an adaptor protein considered necessary for TLR2-induced inflammation, we now report for the first time that oral challenge with P. gingivalis leads to alveolar bone resorption in the absence of MyD88. Indeed, in contrast to prototypical TLR2 agonists, such as the lipopeptide Pam3CSK4 that activates TLR2 in a strictly MyD88-dependent manner, P. gingivalis strikingly induced TLR2 signaling in neutrophils and macrophages regardless of the presence or absence of MyD88. Moreover, genetic or antibody-mediated inactivation of TLR2 completely reduced cytokine production in P. gingivalis-stimulated neutrophils or macrophages, suggesting that TLR2 plays a non-redundant role in the host response to P. gingivalis. In the absence of MyD88, inflammatory TLR2 signaling in P. gingivalis-stimulated neutrophils or macrophages depended upon PI3K. Intriguingly, TLR2-PI3K signaling was also critical to P. gingivalis evasion of killing by macrophages, since their ability to phagocytose this pathogen was reduced in a TLR2 and PI3K-dependent manner. Moreover, within those cells that did phagocytose bacteria, TLR2-PI3K signaling blocked phago-lysosomal maturation, thereby revealing a novel mechanism whereby P. gingivalis can enhance its intracellular survival. Therefore, P. gingivalis uncouples inflammation from bactericidal activity by substituting TLR2-PI3K in place of TLR2-MyD88 signaling. These findings further support the role of P. gingivalis as a keystone pathogen, which manipulates the host inflammatory response in a way that promotes bone loss but not bacterial clearance. Modulation of these host response factors may lead to novel therapeutic approaches to improve outcomes in disease conditions associated with P. gingivalis.

Keywords: MyD88; P. gingivalis; PI3 kinase; TLR2 signaling; immune evasion; macrophages; neutrophils.

Figure 1

Experimental periodontitis induced by oral challenge with P. gingivalis. Groups of mice were administered P. gingivalis in CMC vs. CMC alone by repeated oral gavage. Six weeks later, the maxillae were harvested and alveolar bone volume was measured by μCT from the cemento-enamel junction to a reference line. The residual bone volume of P. gingivalis infected mice was compared to vehicle-treated mice (n = 8–10 per group). (A) WT, Tlr2^−/−, Myd88^−/−, and Tlr2/Myd88 double knock-out mice (DKO) infected with P. gingivalis ATCC 381. (B) An independent experiment examined bone loss in WT vs. Myd88^−/− mice infected with P. gingivalis ATCC 381 vs. P. gingivalis ATCC 53977. Ns, non-significant. **P ≤ 0.01, ***P ≤ 0.005.

Figure 2

Myd88^−/− neutrophils respond to challenge with P. gingivalis. (A) Total Myd88^−/− bone marrow cells and neutrophil-enriched bone marrow cells were plated 4^*10⁵ cells/well in a 96-well plate and incubated with P. gingivalis (MOI 100). (B) Glycogen-induced Myd88^−/− PECs, or naïve bone marrow Myd88^−/− Ly6G+ neutrophils isolated by positive selection (C) produce TNF-α in response to P. gingivalis challenge but not in response to S. minnesota LPS (1 μg/ml) or Pam3CSK4 (10 μg/ml), or buffer control. (D) The response of Myd88^−/− total and neutrophil enriched BM was compared to the response of WT BM cells. (E) BM cells were collected from WT and Myd88^−/− mice and primed with GM-CSF or IFN-γ for 2 h prior to challenge with P. gingivalis (E). The percent increase in TNF produced in response to P. gingivalis challenge in primed cells vs. unprimed cells is shown. (A–E) Supernatants were collected after overnight challenge and the level of TNF in the supernatants was measured by ELISA. One representative experiment is shown in each case. Experiments were repeated 3–5 times. ^*P ≤ 0.05, ^**P ≤ 0.01, ^***P ≤ 0.005.

Figure 3

IFN-γ priming enables TLR2-dependent, MYD88-independent signaling in macrophages in response to P. gingivalis challenge. (A) Naïve Myd88^−/− BMM vs. BMM primed with IFN-γ (100 ng/ml) for 2 h were challenged with P. gingivalis (MOI 100) vs. Pam3CSK4 (10 μg/ml). In (B) Myd88^−/− BMM primed with IFN-γ were challenged with increasing MOI of P. gingivalis. (A,B) Supernatants were collected after overnight stimulation and tested for TNF by ELISA. ^**P ≤ 0.01, ^***P ≤ 0.005.

Figure 4

Kinase involvement in the Myd88^−/− response to P. gingivalis. (A) Ly6C+ BM neutrophils or (B–D) BMM were prepared from Myd88^−/− mice and primed with IFN-γ (100 ng/ml for 2 h). (A,B) Inhibitors for PI3K (LY 2940002 100 μM), p38 MAPK (SB 202190 50 μM), mTORC1 (RaPamycin 60 nM) and RAC1 inhibitor (NSC 23766, 50 μM) were added 30 min before challenge with P. gingivalis (MOI 100). DMSO was used as a control at the highest concentration used in the inhibitor wells. Supernatants were collected after overnight stimulation and the percent inhibition of TNF production is shown. (C) Myd88^−/− BMM were similarly primed and Ly294 was added at increasing concentrations 30 min prior to challenge with P. gingivalis. (D) Myd88^−/− BMM were primed with IFN-γ and antibodies (20 μg/ml anti-TLR2 or TLR4 vs. isotype control, I.C.) or LY294 were added prior to challenge with P. gingivalis. ^*P ≤ 0.05, ^**P ≤ 0.01, ^***P ≤ 0.005.

Figure 5

TLR2-PI3K plays a non-redundant role in the murine and human macrophage response to P. gingivalis. WT murine BMM (A), RAW264.7 macrophages (B), and PMA-differentiated human THP-1 cells (C) were primed with IFN-γ. TLR2 and TLR4 were inhibited with blocking antibodies vs. isotype control (I.C.) for 1 h and PI3K was blocked with LY294 prior to challenge with P. gingivalis (MOI 10). Supernatants were collected after overnight incubation and TNF was measured by ELISA. Background (BG) represents IFN-γ primed cells not challenged with P. gingivalis. Cells challenged with P. gingivalis without any blocker are referred to in the graphs as (–). Representative graphs of >3 repeats are shown. ^*P ≤ 0.05, ^**P ≤ 0.01, ^***P ≤ 0.005.

Figure 6

TLR2-PI3K signaling suppresses phagocytosis and enhances intracellular survival. (A) RAW264.7 or (B) PMA-differentiated THP-1 cells were treated with blocking antibodies or PI3K inhibitor and then challenged with FITC-labeled P. gingivalis at MOI 10 for 1 h. Cells were then washed, extracellular fluorescence was quenched with trypan blue, and phagocytosis was determined using a fluorescence plate reader (RFU, relative fluorescence units). (C,D) RAW 264.7 cells were treated with TLR blocking antibodies or the PI3K inhibitor prior to challenge with P. gingivalis at MOI 10 for 1 h. Cells were then washed and extracellular bacteria were killed by incubating the cells with Metronidazole and Gentamycin for 1 h. Cells were allowed to recover in fresh media for an additional hour after which they were lysed by DDW for 20 min and lysates were plated on blood agar plates in serial dilution. CFU were enumerated after 7 days of anaerobic growth. ^*P ≤ 0.05, ^**P ≤ 0.01, ^***P ≤ 0.005.

Figure 7

TLR2-PI3K signaling enhances intracellular survival by blocking phago-lysosomal maturation. RAW 264.7 cells were seeded at 3 × 104 cells/ well in Ibidi 8 well m-slides. The cells were untreated (A) or treated with anti-TLR2 (B), anti-TLR4 (C), or with the PI3K inhibitor LY 294002 (D) for an hour. Cells were then infected with FITC-labeled P. gingivalis at MOI 10 for 1 h. LysoTracker red was added at 50 nM for the last 10 min of infection. Cells were washed and fixed with 2% formaldehyde and mounted with mounting media. Images were captured using a NIKON confocal microscope at 60X magnification. Yellow color indicates co-localization of P. gingivalis (green) with lysosomes (red). In each field (A–D) the cell in the box is further magnified and shown in the upper right corner. (E) The percent of co-localization was determined by counting cells that demonstrate co-localization as a percentage of all FITC positive cells. (F) Schematic representation of the pathway used by P. gingivalis to evade bactericidal activity without preventing inflammation. ^***P ≤ 0.005.