TLR4 Expression by Liver Resident Cells Mediates the Development of Glucose Intolerance and Insulin Resistance in Experimental Periodontitis

Abstract

Background: Results from epidemiological studies indicate a close association between periodontitis and type 2 diabetes mellitus. However, the mechanism linking periodontitis to glucose intolerance (GI) and insulin resistance (IR) is unknown. We therefore tested the hypothesis that periodontitis induces the development of GI/IR through a liver Toll-like receptor 4 (TLR4) dependent mechanism.

Methods: TLR4 chimeric mice were developed by bone marrow transplantation using green fluorescent protein expressing TLR4WT mouse (GFPWT) as donor and TLR4 WT or TLR4-/- as recipient mice (GFPWT:WT and GFPWT:KO chimeras respectively). These chimeras were subjected to experimental chronic periodontitis induced by repeated applications of LPS to the gingival sulci for 18 weeks. The levels of GI/IR were monitored and plasma cytokines and LPS were determined at 18 weeks when differences in glucose tolerance were most apparent. Cytokine gene expression was measured in liver tissue by qPCR.

Results: Alveolar bone loss was significantly greater in GFPWT:WT chimeras treated with LPS compared with chimeras treated with PBS or GFPWT:KO chimeras. However, the degree of gingival inflammation was similar between GFPWT:WT and GFPWT:KO mice with LPS application. Severe GI/IR occurred in GFPWT:WT chimeras but not in the GFPWT:KO chimeras that were subjected to 18 weeks of LPS. Serum LPS was detected only in animals to which LPS was applied and the level was similar in GFPWT:WT and GFPWT:KO mice at the 18 week time point. Surprisingly, there was no significant difference in the plasma levels of IL1β, IL6 and TNFα at 18 weeks in spite of the severe GI/IR in the GFPWT:WT chimeras with LPS application. Also, no difference in the expression of TNFα or IL6 mRNA was detected in the liver of GFPWT:WT vs GFPWT:KO mice. In contrast, liver IL1β expression was significantly greater in GFPWT:WT chimeras compared to GFPWT:KO chimeras treated with LPS.

Conclusion: We observed that GFPWT:WT, but not GFPWT:KO chimeras, treated with LPS developed GI/IR despite similar degrees of gingival inflammation, circulating cytokine levels, and LPS concentrations. We conclude that LPS from periodontitis sites has a pivotal role in triggering the development of GI/IR through a mechanism that involves TLR4 expression by resident macrophages/Kupffer cells in the liver.

Fig 1. Hypothetical impact of LPS from the periodontium on the liver following LPS application to the gingival sulcus and the effects on glucose tolerance.

Pathways represent the proposed activity in the periodontium and in the liver in GFPWT:WT (A) and GFP-WT:KO chimeras (B) following LPS application to the gingival sulcus. (A) LPS from gram negative bacteria is released in the gingival sulci where it acts as a chemoattractant. Macrophages (TLR4+/+) from circulating blood migrate to the gingival tissues where they are activated by LPS and secrete proinflammatory cytokines such as TNFα, IL1β and IL6. These cytokines and/or LPS enter the circulation and eventually reach the liver. Liver Kupffer cells that reside in sinusoids will be exposed to either LPS and/or proinflammatory cytokines. In GFPWT:WT animals Kupffer cells are TLR4+/+ and will respond to LPS by synthesizing and secreting proinflammatory cytokines resulting in impaired ipGTT. In addition, cytokines from the circulation can impair insulin signaling pathway in hepatocytes. (B) In GFPWT:KO animals, Kupffer cells are TLR4-/- and if the major stimulus to the liver Kupffer cells is LPS, ipGTT in these animals will be normal. However, circulating proinflammatory cytokines may act directly on hepatocytes leading to impaired ipGTT. IR: insulin resistance, ipGTT: intraperitoneal glucose tolerance test.

Fig 2. Schematic illustration of the experimental timeline and generation of chimeric mice.

(A) Timeline, (B) Generation of chimeric mice: Bone marrow derived cells (BMDC) were isolated from a donor mouse (GFPWT) and transplanted to an irradiated recipient mouse (either WT or TLR4-/-). Following 5 weeks of consumption of water with antibiotics (AB), peripheral blood was drawn and FACS analysis performed confirming that >90% of circulating blood cells expressed GFP.

Fig 3. Further confirmation of chimera generation by immunofluorescence microscopy and qPCR.

Donor macrophages migrate into the gingiva of both GFPWT:WT and GFPWT:KO mice, while Kupffer cells retain their recipient phenotype. (A) Immunofluorescence microscopy identifying TLR4+ macrophages (ED1+) in gingival tissue adjacent to a tooth in GFPWT:WT (top panel) and GFPWT:KO (bottom panel) chimeras following LPS application. Gingival macrophages in both GFPWT:WT and GFPWT:KO chimeras express TLR4+, GFP+, and ED1+ following LPS application. Red: TLR4+, Green: GFP+, Blue: ED1+. Arrows indicate migrating donor macrophages. (B) Immunofluorescence microscopy identifying liver Kupffer cells maintaining recipient’s phonotype in a GFPWT:WT chimera following LPS application (top panel, arrows: TLR4+, GFP-, ED1+ cells). Kupffer cells in GFPWT:KO are TLR4-, GFP-, ED1+ (bottom panel, arrows: TLR4-, GFP-, ED1+ cells). The few GFP+,TLR4+, ED1+ cells are likely migrating donor macrophages replacing recipient Kupffer cells. Red: TLR4+, Green: GFP+, Blue: ED1+. (C) Real time PCR analysis of liver TLR4 (a) and F4/80 (b) expression. Liver expression of TLR4 and the macrophage marker F4/80 was analyzed by TaqMan Real Time PCR. TLR4 expression was significantly higher in GFPWT:WT animals compared to GFPWT:KO mice (a). There is no statistical difference in F4/80 expression between GFPWT:WT and GFPWT:KO mice. x-axis: chimeric group, y-axis: relative expression of TLR4 (a) and F4/80 (b). All normalized to β-actin. n = 8–12 per chimeric group (4–6 per treatment group). Data presented are mean ± SD. *p<0.05, **p<0.001.

Fig 4. Serum LPS concentration is determined in animals with LPS vs PBS application.

LPS is present only in the plasma of animals treated with LPS. x-axis: chimeric group, y-axis: LPS concentration (EU/ml). Data presented are mean ± SD. n = 8–12 per chimeric group (4–6 per treatment group). Data presented are mean ± SD. *p<0.001

Fig 5. LPS application to gingival sulci induced gingival inflammation and bone loss.

Gingival inflammation was present in both chimeras after gingival application of LPS (A-F). The degree of gingival inflammation was confirmed by the number of inflammatory infiltrates (A) and identification of IL1β (B), TNFα (C), IL6 (D), MCP1 (E), and MMP9 (F) producing cells. Inflammatory infiltrates were quantitatively measured using sections stained with H&E (n = 4 per treatment group) (A). Number of cells expressing cytokines (B, C, D), MCP1(E), and MMP9 (F) was assessed by immunofluorescence microscopy (n = 4 per treatment group). Inflammatory infiltrates were present in both GFPWT:WT and GFPWT:KO mice treated with LPS (A). Insets show inflammatory cell infiltrates. Arrows point to inflammatory cells. The number of inflammatory cells was significantly higher in LPS treated animals compared to PBS treated animals in both chimeric groups (p< 0.01) (A). There was no statistical difference in the amount of infiltrates between these chimeras with LPS application (A). There was also no statistical difference in the number of cells stained with IL1β, TNFα, IL6, MCP1 and MMP9 (B-F) between GFPWT:WT and GFPWT:KO mice with LPS application. Arrows indicate GFP+ donor cells expressing IL1β (B), TNFa (C), IL6 (D), MCP1 (E), and MMP9 (F) respectively. Inflammatory cell infiltrates were counted in 5 random fields visualized by light microscopy at 400X magnification. Photos of H&E sections were taken at 200X magnification. Insets are at 400X (A). Cells positive for cytokines, MCP1 and MMP9 were counted in 5 random fields and visualized by immunofluorescence microscopy at 400X magnification. Photos of immunofluorescence microscopy were taken at 400X (B-F). In all cases, the examiner was blinded to the animal group to which the tissue section belonged. x-axis: chimeric group. y-axis: number of cells/random field. Data presented are mean ± SD. (G) Alveolar bone loss was assessed by microtomography: Microtomography of maxilla from GFPWT:WT chimeras following 18 weeks of LPS or PBS treatment. The distances between the cemento-enamel junction (CEJ) and the alveolar crest are shown in red. The distance between the CEJ and the alveolar bone shown in a GFPWT:WT chimera without LPS application is occupied by junctional epithelium and connective tissue attached to the root and therefore does not reflect actual bone loss (G-a). Comparison of bone loss among chimeric groups (G-b):The bracket on the right bottom indicates the mean distance where epithelia and connective tissue attach. The bracket on the top indicates the mean bone loss. Bone loss was statistically greater in GFPWT:WT mice with LPS compared to PBS treatment. There is no statistically significant difference in bone loss between PBS and LPS treated groups in GFPWT:KO animals (G-b). x-axis: chimeric group, y-axis: distance (mm) measured from CEJ to the crest of alveolar bone on the mid palatal portion of the second maxillary molars. n = 8–12 per chimeric group (4–6 per treatment group). Data presented are mean ± SD. *p<0.05.

Fig 6. Glucose tolerance test results over 18 weeks.

Severe glucose intolerance developed by 18^th week in GFPWT:WT animals with periodontitis. ipGTT was performed at week 0 (baseline), 5, 10, 14 and 18 weeks in GFPWT:WT (upper panel) and GFPWT:KO chimeras (lower panel) with LPS or PBS application. There was no difference in ipGTT results between these chimeras at baseline (data not shown). Data presented at each time point are mean ± SD. x-axis: post dextrose injection in min, y-axis: glucose levels (mg/dL). n = 8–12 per chimeric group (4–6 per treatment group). *p< 0.05, **p< 0.01, ***p< 0.001.

Fig 7. Fasting glucose, insulin and HOMA-IR measured over 18 weeks.

Fasting glucose (A), insulin (B) and HOMA-IR (C) were measured at baseline and week 5, 10, 14 and 18. IR and fasting insulin levels were significantly higher in GFPWT:WT animals with LPS compared to PBS application at 18 weeks. x-axis: chimeric groups; GFPWT:WT, GFPWTKO, C: control (PBS application), P: periodontitis (LPS application). y-axis: fasting glucose concentration (mg/dL) (A), fasting insulin concentration(μg/L) (B), and HOMA-IR index (C). Data presented are mean ± SD. n = 8–12 per chimeric group (4–6 per treatment group). * p<0.05.

Fig 8. Plasma levels of cytokines measured by ELISA.

Plasma levels of IL1β, TNFα, and IL6 which are known to cause IR were measured by Q-Plex ELISA. x-axis: chimeric group. y-axis: concentration in pg/ml. Data at 18 weeks presented are mean ± SD. There was no statistical difference among groups with either LPS or PBS application (p>0.05), n = 8–12 per chimeric group (4–6 per treatment group).

Fig 9. Expression of IL1β and TNFα in the liver determined by qPCR.

(A) There was a significantly higher level of IL1β in GFPWT:WT mice treated with LPS compared to PBS treated animals and GFPWT:KO mice in both LPS and PBS treated groups. There was no significant increase in IL1β in animals treated with LPS or PBS in GFPWT:KO animals. (B) There was no significant difference in TNFα expression between chimeras with and without LPS application. IL-6 was not detected in the liver (data not shown). x-axis: chimeric group, y-axis: relative gene expression levels. Data presented are mean ± SD. n = 8–12 per chimeric group (4–6 per treatment group). *p<0.05.