Endocytic pathways regulate Toll-like receptor 4 signaling and link innate and adaptive immunity

Abstract

Immune responses are initiated when molecules of microbial origin are sensed by the Toll-like receptors (TLRs). We now report the identification of essential molecular components for the trafficking of the lipopolysaccharide (LPS) receptor complex. LPS was endocytosed by a receptor-mediated mechanism dependent on dynamin and clathrin and colocalized with TLR4 on early/sorting endosomes. TLR4 was ubiquitinated and associated with the ubiquitin-binding endosomal sorting protein hepatocyte growth factor-regulated tyrosine kinase substrate, Hrs. Inhibition of endocytosis and endosomal sorting increased LPS signaling. Finally, the LPS receptor complex was sorted to late endosomes/lysosomes for degradation and loading of associated antigens onto HLA class II molecules for presentation to CD4+ T cells. Our results show that endosomal trafficking of the LPS receptor complex is essential for signal termination and LPS-associated antigen presentation, thus controlling both innate and adaptive immunity through TLR4.

Figure 1

The LPS receptor complex is endocytosed and appears on endosomal structures in human monocytes. (A) Flow cytometry showing downregulation of surface TLR4. The surface TLR4 level was monitored in monocytes before and after LPS (2 ng/ml) stimulation using the TLR4 antibody HTA125 and the mouse IgG2a as an isotype control. (B) Confocal images of monocytes stimulated with Cy5-labeled LPS (LPS^CY5). The monocytes were incubated with LPS^Cy5 (250 ng/ml) and monitored over a period of 5 h. Arrows indicate LPS-positive endosomal structures. (C) Confocal images of monocytes stimulated with OregonGreen-labeled LPS (LPS^OG) and coincubated with Alexa546-labeled HTA125 (anti-TLR4^A546). Images were captured 1 h after addition of anti-TLR4^A546 (2 μg/ml) and LPS^OG (600 ng/ml). No uptake was observed for the Alexa546-labeled mouse isotype control IgG2a (2 μg/ml) (data not shown). Bar=5 μm.

Figure 2

LPS is internalized on early endosomes containing TLR4. HEK293 cells expressing TLR4^YFP/MD-2 or TLR4^YFP only were transfected with CD14 and stimulated by adding LPS^Cy5 (250 ng/ml). (A) Confocal images before (left) and 30 min after (right) the addition of LPS^Cy5. (B) Estimation of the number TLR4^YFP-containing endosomes (>0.5 μm) before and after stimulation. The TLR4^YFP-containing endosomes were counted in living cells without stimulation and in the period 10–40 min after LPS^Cy5 stimulation. The number of cells analyzed were n=52 cells for each situation. (C) Comparison of the number of TLR4^YFP-containing endosomes (>0.5 μm) with and without coexpression of MD-2. The LPS-positive endosomes with TLR4 colocalization was quantified as above in CD14-transfected cells expressing TLR4^YFP/MD-2 (n=85) or cells expressing TLR4^YFP (n=103). (D) Confocal images of cells showing colocalization of TLR4 and LPS on early/sorting endosomes. HEK293 cells expressing TLR4^YFP/MD-2 cotransfected with CD14 and the marker for early/sorting endosomes, 2 × FYVE^CFP, 45 min after the addition of LPS^Cy5. (E) Confocal image of a cell transfected as in (D) and simultaneously given Alexa⁵⁴⁶-labeled transferrin Tf^A546 (5 μg/ml) and LPS^Cy5. Full-sized images of each of the four channels of the overlaid image are shown in Supplementary Figure 2. All the images presented in this figure represent optical sections of 1.3 μm and one individual scan per cell was used for the counting of TLR4-positive endosomes. Bar=5 μm.

Figure 3

Endocytosis of the LPS receptor complex is dependent on dynamin and clathrin. (A) Confocal images of dynamin WT (left)- and Dyn K44A (right)-transfected cells. HEK293 cells expressing TLR4^YFP/MD-2 were transfected with CD14 together with dynamin WT or Dyn K44A and left for 6 days before analysis. The cells were simultaneously given LPS^Cy5 (250 ng/ml) and Alexa⁵⁴⁶-labeled transferrin (Tf^A546) (5 μg/ml) and imaged after 50 min. (B) Comparison of the total LPS uptake in dynamin WT (n=46)- and Dyn K44A (n=52)-transfected cells in fixed cells 75 min after LPS^Cy5 stimulation. The LPS uptake was quantified using the LSM510 software by measuring total intracellular fluorescence in manually drawn regions of interest and the number of cells analyzed were n=46 for the dynamin WT and n=52 for the Dyn K44A transfectants. (C) Confocal images of control siRNA-treated cells (left) and clathrin siRNA-treated cells (right) following LPS^Cy5 stimulation. HEK293 cells expressing TLR4^YFP/MD-2 were treated with control or clathrin siRNA for 48 h followed by a 24 h transfection with CD14. The cells were simultaneously given LPS^Cy5 (250 ng/ml) and Tf^A546 (5 μg/ml) and imaged after 35 min. (D) Comparison of the number of LPS-positive endosomes (>0.5 μm) per cell in cells transfected with control siRNA and clathrin siRNA 10–40 min after LPS^Cy5 stimulation. The number of cells monitored was n=33 (control siRNA) and n=25 (clathrin siRNA). Clathrin- and control siRNA-treated HEK293 cells transfected with CD14 alone were included for comparison (n=40 and 16, respectively). Bar=10 μm.

Figure 4

LPS promotes TLR4 ubiquitination and the association of TLR4 with Hrs. (A) HEK293 cells expressing TLR4^YFP/MD-2 were stimulated with LPS (1 μg/ml) for the indicated time periods. TLR4^YFP was immunoprecipitated with a polyclonal anti-GFP antibody and blotted for ubiquitin (FK2 antibody) and GFP (monoclonal GFP antibody). (B) Monocytes were either left untreated or stimulated with LPS (1 μg/ml) for the indicated time periods. Cellular lysates were made and TLR4 immunoprecipitated with a polyclonal TLR4 antibody and blotted for ubiquitin (left), TLR4 (middle) and Hrs (lower left) or the reverse immunoprecipitatation of Hrs or control IgG blotted for TLR4 (upper right). The Hrs in lysate is shown in the lower right panel. (C) HEK293 cells expressing TLR4^YFP/MD-2 were transfected with c-myc Hrs for 24 h before LPS stimulation for the indicated periods. Cellular lysates were prepared and immunoprecipitated with anti-GFP (TLR4^YFP) or control IgG (C1) and blotted for Hrs. Control C2 represents Hrs immunoprecipitated from nontransfected HEK293 cells showing the size of endogenous Hrs. Overexpression of Hrs resulted in an LPS-induced reduction in TLR4 levels in lysates, whereas Hrs and tubulin levels are unaffected (lower three panels). (D) Confocal images of HEK293 cells expressing TLR4^CFP/MD-2 transfected with CD14 and Hrs^YFP following 3 h of LPS (250 ng/ml) stimulation. Bar=5 μm.

Figure 5

Lysosomal degradation of TLR4 limits signaling. (A) HEK293 cells expressing TLR4YFP/MD-2 were transfected with CD14 and treated with chloroquine (10 μM) or lactacystin (10 μM) for 30 min prior to LPS (1 μg/ml) stimulation for the indicated time periods. Cellular lysates were made and TLR4^YFP was detected using a polyclonal anti-GFP antibody and reblotted for α-tubulin as a control for equal loading. The samples displayed represent sections of the original blot. (B) HEK293 cells expressing TLR4^YFP/MD-2 were transfected with CD14 and treated with 10 μM LY or control vehicle (EtOH) for 30 min and stimulated with LPS (1 μg/ml) for 10 h. Cellular lysates were made and TLR4^YFP was detected using a polyclonal anti-GFP antibody and reblotted for α-tubulin as a control for equal loading. (C) HEK293 cells expressing TLR4^YFP/MD-2 were treated with control of Hrs siRNA and transfected with CD14. Cellular lysates were made and blotted for GFP (TLR4^YFP), Hrs, and α-tubulin for the examination of TLR4^YFP and Hrs levels. The samples displayed represent sections of the original blot. (D) Effect of dynamin and ubiquitination on LPS-induced NFκB activation through TLR4. HEK293 cells expressing TLR4 were transfected with MD-2, CD14 and Elam-luc together with WT constructs (dynamin and UbRGG) or a dominant negative version of dynamin (Dyn K44A) or an inhibitory version of ubiquitin (UbR, lacking the two C-terminal glycines necessary for conjugation to proteins but able to occupy ubiquitin-binding pockets). The cells were stimulated for 6 h with LPS (10 ng/ml) and assayed for NFκB-induced luciferase expression. (E) Effect of Hrs and ESCRT-I on LPS-induced NFκB activation through TLR4. HEK293 cells expressing TLR4^YFP/MD-2 were treated with control siRNA or siRNA directed towards Hrs or HCRP1 (component of ESCRT-I) and cotransfected with CD14 and Elam-luc as described in the Materials and methods section and assayed for NFκB-induced luciferase expression.

Figure 6

The LPS receptor complex is trafficked to the class II loading compartment. (A) Confocal images showing TLR4^YFP in the lumen of LAMP-1-positive endosomes. HEK293 cells expressing TLR4^YFP/MD-2 transfected with CD14 were treated for 2 h with LPS (1 μg/ml) and fixed before intracellular staining for LAMP1. Arrows indicate LAMP-1-positive endosomes showing intraluminal TLR4^YFP. (B) Confocal images showing TLR4^YFP in the lumen of nonacidic endosomes. HEK293 cells expressing TLR4^YFP/MD-2 transfected with CD14 were stimulated with LPS (1 μg/ml) for 3 h and incubated with 10 nM LysoTracker Red (C). Antibodies towards TLR4 (HTA125), CD14 (5C5) or MD-2 (IIC1) or isotype-matched controls were added to human HLA-matched monocytes together with cloned DR4-restricted CD4⁺ T cells specific for mouse Cκ^40–48. T-cell proliferation was measured after 48 h by incorporation of [³H] TdR. Data points are mean of duplicates from a representative experiment. Isotype-matched monoclonal antibodies were used as controls. Bar=5 μm.