Raftlin Controls Lipopolysaccharide-Induced TLR4 Internalization and TICAM-1 Signaling in a Cell Type–Specific Manner

The transportation of microbial components from an extracellular milieu to the intracellular compartments is essential for the activation and signaling of endosomal TLRs (1). TLR4 is a cell-surface type I transmembrane protein that recognizes LPS cooperatively with MD-2 (2, 3). During Gram-negative bacterial infection, extracellular LPS is sensed by LPS binding protein and CD14, which helps the TLR4–MD-2 complex bind LPS (4). LPS-induced TLR4-mediated signaling is compartmentalized into the plasma membrane and endosomes, where proinflammatory cytokine and IFN-β production are induced, respectively (1). The LPS-bound TLR4 undergoes dimerization and associates with the sorting adaptor TIRAP (also called Mal) that localizes to the phosphatidylinositol-(4, 5)-bisphosphate–rich plasma membrane (5). TIRAP-dependent binding of MyD88 recruits the signaling molecules IL-1R–associated kinases 1 and 4, which forms so-called myddosomes that activate NF-κB and MAPK signaling cascades, leading to production of proinflammatory cytokines (6, 7). While TIRAP-MyD88–dependent signaling occurs at the plasma membrane, Toll–IL-1R domain–containing adaptor molecule-2 (TICAM-2; also called TRAM)– and TICAM-1 (also called TRIF)–dependent IFN-regulatory factor 3 (IRF3) activation initiates from the endosomes (8–14). Dissociation from TIRAP-MyD88 and subsequent clathrin- and dynamin-mediated internalization of LPS-TLR4 are prerequisites for binding to the sorting adaptor TICAM-2 that localizes to the endosomal membrane via myristoylation at the N terminus (12, 15). TLR4-mediated TICAM-2 dimerization on the endosomal membrane induces TICAM-1 recruitment and oligomerization that led to IRF3 and NF-κB activation, resulting in IFN-β production. Membrane-bound CD14 mediates the LPS-induced endocytosis of TLR4 and the tyrosine kinase Syk regulates this process (16). Moreover, the p110δ isoform of PI3K acts as a regulator of TLR4 endocytosis in dendritic cells (DCs) through modulating phosphatidylinositol-(4, 5)-bisphosphate metabolism (17). However, the precise molecular mechanisms that control the translocation of LPS-TLR4 and spatiotemporal regulation of TLR4-mediated signaling are still unknown.

Recently, we have demonstrated that the cytoplasmic protein Raftlin is indispensable for cell entry of extracellular polyinosinic:polycytidylic acid [poly(I:C)], a virus dsRNA analogue, and TLR3–TICAM-1–mediated signaling (18). Raftlin was originally identified as a major raft protein with molecular mass of 63 kDa in B cells that colocalized with BCR in the lipid raft before and after BCR activation (19). Raftlin also localizes to the lipid raft in T cells and modulates BCR and TCR signaling (20). Although Raftlin possesses fatty acylation sites at the N terminus, it predominantly localizes in the cytoplasm in human epithelial cells and monocyte-derived DCs (Mo-DCs) (18, 19). Upon poly(I:C) binding to the uptake receptor on the cell surface, Raftlin moves from the cytoplasm to the plasma membrane, where it induces endocytosis of the uptake receptor via interaction with the clathrin–clathrin-associated adaptor protein–2 (AP-2) complex. Raftlin also mediates the internalization of B- and C-type CpG or control GpC oligonucleotides that share the uptake receptor with poly(I:C), but it is not involved in the clathrin-mediated endocytosis of transferrin (18, 21). Thus, Raftlin appears to select cargo derived from microbes and modulates innate immune activation independently of raft-mediated signal modulation. In this study, we showed that Raftlin is required for LPS-induced TLR4 internalization and TICAM-1–mediated IFN-β production in human Mo-DCs and monocyte-derived macrophages (Mo-Mϕs).

HEK293 cells were maintained in DMEM low glucose (Invitrogen) supplemented with 10% heat-inactivated FCS (Thermo Scientific) and penicillin/streptomycin. MRC5 cells were maintained in MEM-α (Invitrogen) supplemented with 5% heat-inactivated FCS and penicillin/streptomycin. HEK293-TLR4/MD2/CD14 cells were purchased from Invivogen and maintained in DMEM high glucose supplemented with 10% heat-inactivated FCS, penicillin/streptomycin, 100 μg/ml normocin, and 2 mM l-glutamine. HeLa cells were maintained in MEM (Nissui, Tokyo, Japan) supplemented with 1% l-glutamine and 5% heat-inactivated FCS. The anti-human Raftlin rabbit serum and anti-human TLR4 mouse mAb (HTA125) were provided by Dr. K. Saeki (Juntendo University) and Dr. K. Miyake (University of Tokyo), respectively. The anti-Raftlin goat polyclonal Ab (pAb) (N-14), anti-Rab5a rabbit pAb, and anti-clathrin HC mAb (TD.1) were purchased from Santa Cruz Biotechnology; anti-clathrin HC mAb (x-22) was from Abcam; anti–β-actin mAb (A2228) was from Sigma; anti-adaptin β mAb was from BD Transduction Laboratories; allophycocyanin-labeled anti-mouse TLR4 rat mAb (SA15-21), PE-labeled anti-human TLR4 mAb (HTA125), and PE-labeled anti-human CD14 mAb (M5E2) were from BioLegend; FITC-labeled anti-mouse CD14 rat mAb (Sa2-8) and allophycocyanin-, PE-, or FITC-labeled isotype-matched control Ab were from eBioscience; Alexa Fluor–conjugated secondary Abs were from Invitrogen. The LPS from Escherichia coli (0111:B4), chlorpromazine, methyl-β-cyclodextrin, and Dynasore were purchased from Sigma. Poly(I:C) was from GE Bioscience. Pam2CSK4 was synthesized by Biologica (Nagoya, Japan).

The expression plasmids for human TLR4 (pEF-BOS/TLR4) and MD-2 (pEF-BOS/MD-2) were provided by Dr. K. Miyake (University of Tokyo). The pME18S/CD14 expression plasmid was provided by Dr. H. Nishimura (University of Tsukuba). The Gal4-IRF3 and p55 UASG-Luc plasmids were provided from Dr. T. Fujita (Kyoto University).

A C57BL/6 background Rftn2-targeting vector was purchased from the European Conditional Mouse Mutagenesis Program. After the targeting vector was transfected into C57BL/B6N-derived embryonic stem cells (22), G418-resistant colonies were selected and screened by PCR. Chimeric mice were generated by aggregation of the mutated embryonic stem cells at the eight-cell stage. To remove exon 2 of Rftn2, we crossed the Rftn2 heterozygous mutants with Cre-transgenic mice. Rftn^−/− mice were provided by Dr. A. Yoshimura (Keio University). The Rftn2 homozygous mutants were crossed with Rftn^−/− mice to generate Rftn/Rftn2 double-knockout mice. Inbred C57BL/6 wild-type mice were purchased from CLEA Japan (Tokyo, Japan). Mice were maintained under specific pathogen-free conditions in the animal facility of the Hokkaido University Graduate School of Medicine. Female mice aged 8–14 wk were used in all experiments, which were performed according to the guidelines established by the Hokkaido University Animal Care and Use Committee.

HEK293 cells (2 × 10⁵ cells/well) cultured in 24-well plates were transfected with the control small interfering RNA (siRNA) or siRNA for RFTN (5 pmol) using Lipofectamine RNAiMAX (Invitrogen). Twenty-four hours after siRNA transfection, cells were washed, reseeded in 96-well plates, and then transfected with indicated expression vectors together with the reporter plasmid (25 ng/well) and an internal control vector, phRLTK (Promega) (1.25 ng/well) using Lipofectamine 2000 (Invitrogen). The p-125 luciferase reporter that contained the human IFN-β promoter was provided by Dr. T. Taniguchi (University of Tokyo). A luciferase-linked NF-κB reporter gene was from Stratagene. The reporter plasmid containing the endothelial leukocyte adhesion molecule-1 promoter was constructed in our laboratory. The Gal4-IRF3 and p55 UASG-Luc for IRF3 activation were described previously (23). The total amount of DNA (100 ng/well) was kept constant by adding empty vector. After 24 h, cells were stimulated with LPS for 6 h or left untreated. Thereafter, cells were lysed in lysis buffer (Promega), and firefly and Renilla luciferase activities were determined using a Dual-Luciferase reporter assay kit (Promega). The firefly luciferase activity was normalized to the Renilla activity and expressed as the fold-stimulation relative to the activity of unstimulated cells. All assays were performed in triplicate.

siRNA duplexes (RFTN, catalog no. s23219; negative control, catalog no. AM4635) were obtained from Ambion–Applied Biosystems. HEK293 cells cultured in 24-well plates were transfected with siRNA (5 pmol) using Lipofectamine RNAiMAX (Invitrogen). Forty-eight hours after transfection, total RNA was extracted by TRIzol reagent (Invitrogen), or cells were lysed with lysis buffer including 1% Nonidet P-40 (NP-40) for quantitative PCR or immunoblotting, respectively, to assess the knockdown efficiency. Human monocytes were cultured on microcover glasses (Matsunami, Tokyo, Japan) in 24-well plates with 20 ng/ml human GM-CSF. At days 4 and 5, cells were transfected with 10 pmol siRNA using Lipofectamine RNAiMAX. At day 6, cells were fixed and stained for confocal microscopy.

Lentiviral vectors expressing control short hairpin RNA (shRNA) or shRNA for Raftlin were constructed using BLOCK-iT U6 RNAi Entry Vector kit (Invitrogen) and BLOCK-iT RNAi Expression System (Invitrogen). Sequences of oligonucleotides for construction of pLenti vector expressing shRFTN were as follows: shRFTN (top 5′-CACCGCTTGACAGATGGAGTATTCACGAATGAATACTCCATCTGTCAAGC-3′, bottom 5′-AAAAGCTTGACAGATGGAGTATTCATTCGTGAATACTCCATCTGTCAAGC-3′). Prepared pLenti vectors were transfected into HEK293 FT cells with pLP1, pLP2, and pLP/VSVG, and the virus particles were concentrated by centrifugation at 20,000 rpm at 4°C for 2 h. The precipitate was resuspended in RPMI 1640 as the virus 1000× concentrated solution. At day 5 of macrophage or DC induction from human CD14⁺ monocytes, lentivirus concentrated solution (1:100) was added to the cells together with Polybrene (1 μg/well) and centrifuged at 660 × g for 90 min. Forty-eight hours after replacement with fresh medium, cells were stimulated with LPS for 24 h. The concentration of IFN-β and IL-6 in collected supernatants was measured using Human IFN-β ELISA kit (TFB) and Human IL-6 ELISA kit (R&D Systems), respectively. To check the knockdown efficiency, we extracted total RNA.

HEK293-TLR4/MD2/CD14 cells were stimulated with LPS for 10 or 30 min or left untreated and lysed in lysis buffer (20 mM Tris-HCl, pH 7.5, containing 150 mM NaCl, 1% NP-40, 10 mM EDTA, 2 mM PMSF, 5 mM Na₃VO₄, and a protease inhibitor mixture; Roche Diagnostics). Lysates were precleared with protein G–Sepharose (GE Healthcare) and incubated with control rabbit serum, control mouse IgG, anti-Raftlin rabbit serum (1:200), anti-clathrin mAb (x-22), or anti–AP-2 mAb. Immune complexes were recovered by incubation with protein G–Sepharose, washed four times with wash buffer (20 mM Tris-HCl, pH 7.5, containing 100 mM NaCl, 0.5% NP-40, 10 mM EDTA, and 10% glycerol), and resuspended in denaturing buffer. Samples were analyzed by 7.5% SDS-PAGE under reducing conditions followed by immunoblotting with indicated Abs and HRP-labeled secondary Abs.

Total RNA was extracted using TRIzol reagent and reverse transcribed using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems) and random primers. Quantitative PCR was performed using the indicated primers (Supplemental Table I) and the Step One Real-time PCR system (Applied Biosystems).

Human monocytes (5.0 × 10⁵/well) were cultured on microcover glasses in a 24-well plate with 20 ng/ml human GM-CSF. At day 6, cells were stimulated with LPS and fixed with 4% paraformaldehyde at indicated time points. After permeabilization with PBS containing 0.02% Triton X-100 and 1% BSA, cells were blocked with PBS containing 1% BSA and 10% goat serum and labeled with anti-Raftlin pAb, anti-Rab5a pAb, anti-TLR4 mAb (1/200), or isotype-matched Abs for 1 h at room temperature. Then Alexa Fluor 488– or 594–conjugated secondary Abs (1/400) were used to visualize the primary Abs. Cells were mounted with ProLong Gold antifade reagent with DAPI (Invitrogen) and visualized at 63× magnification using an LSM510 META microscope (Zeiss, Jena, Germany). Colocalization coefficients were calculated using LSM510 ZEISS LSM Image examiner software (24).

Cells were washed once with FACS buffer (Dulbecco’s PBS containing 0.5% BSA and 0.1% sodium azide), incubated with Fc receptor blocking reagent for 5 min on ice, and then reacted with indicated mAbs or isotype-matched control Ab for 30 min at 4°C. After washing twice with FACS buffer, cells were analyzed on a FACSCalibur (BD Biosciences). The flow cytometry data were analyzed by FlowJo software.

Immunoprecipitated samples were separated on a 10% SDS-PAGE under reducing conditions, and the region of the gels containing proteins from ∼250 to 20 kDa was cut at 1- to 2-mm intervals as described previously (25). After in-gel digestion with modified trypsin, the resulting peptides were analyzed by liquid chromatography tandem mass spectrometry. The ion spectrum data generated by liquid chromatography tandem mass spectrometry were screened against the international protein index human database (version 3.29) with Mascot (Matrix Science, London, U.K.) to identify high-scoring proteins.

Splenocytes from wild-type or knockout mice were treated with 400 IU MandleU per milliliter collagenase D (Roche) at 37°C for 25 min in HBSS (Sigma-Aldrich). EDTA was added to the cell suspensions and incubated for an additional 5 min at 37°C. After lysis of RBCs using the ACK lysis buffer, splenocytes were incubated with MACS anti-CD11c–conjugated microbeads, and DCs were purified using magnetic separation columns as indicated by the manufacturer (Miltenyi Biotec). Positively selected cells were isolated and suspended in RPMI 1640 (Invitrogen) supplemented with 10% heat-inactivated FCS and antibiotics. Bone marrow (BM) cells were prepared from the femur and tibia and cultured in RPMI 1640 supplemented with 10% FCS and 30% culture supernatant of L929 cells expressing M-CSF. At days 3 and 5, cells were washed with PBS and fresh medium was added. At day 6, cells were collected as BM-derived macrophages (BM-Mϕs).

Cells were stimulated with the TLR ligands. In the case of inhibitor treatment, cells were preincubated with chlorpromazine (10, 25 μg/ml), methyl-β-cyclodextrin (0.5, 1.0 mM), Dynasore (50, 100, 200 μM), or 0.1% DMSO for 1 h and stimulated with 100 ng/ml LPS. Twenty-four hours after stimulation, culture supernatants were collected and analyzed for cytokine levels with ELISA or cytometric bead array (CBA). ELISA kit for IFN-β was purchased from PBL Biomedical Laboratories (for mouse IFN-β) and TFB (for human IFN-β). CBA flex sets for human TNF-α, mouse IL-6, and mouse TNF-α were purchased from BD Biosciences. Experiments were performed according to the manufacturer’s instructions, and samples were analyzed using FACSAria (BD Biosciences).

Statistical analyses were made with the Student t test. The p value of significant differences was noted.

The requirement of Raftlin in TLR4 signaling was examined using Raftlin knockdown HEK293 cells transiently expressing TLR4 and MD-2. LPS-induced TLR4-mediated IFN-β promoter activation was significantly decreased in Raftlin knockdown cells compared with control cells, while NF-κB activation was unaffected by Raftlin knockdown (Fig. 1A). LPS-induced TLR4–TICAM-1–mediated IRF3 activation was also abolished by knockdown of Raftlin (Fig. 1B). Because membrane-bound CD14 is critical for TLR4 internalization and LPS-induced IFN-β production in mouse macrophages and DCs (16), we investigated whether the function of Raftlin was affected by CD14. Ectopic expression of CD14 enhanced the activity of TLR4 in response to LPS, and Raftlin knockdown significantly decreased LPS-induced TLR4-mediated IFN-β promoter activation (Fig. 1C), suggesting that Raftlin regulates TLR4–TICAM-1 signaling either in the presence or the absence of CD14 in HEK293 cells. Fig. 1D shows efficacy of Raftlin knockdown by siRNA.

The necessity of Raftlin in LPS-induced IFN-β production was examined using human primary immune cells. Mo-DCs and Mo-Mϕs expressed Raftlin but not Raftlin-2 mRNA (Fig. 2A) (18). To assess the requirement of Raftlin in LPS-induced IFN-β production, we knocked down Raftlin in Mo-DCs and Mo-Mϕs by infection with Raftlin-shRNA–expressing lentiviral particles and analyzed the cellular response to LPS (Fig. 2B). Raftlin expression was decreased in Mo-DCs and Mo-Mϕs (Fig. 2B, right panels). LPS-induced IFN-β production was significantly decreased in Raftlin knockdown Mo-DCs and Mo-Mϕs compared with control cells (Fig. 2B, left panels). IL-6 production was partially decreased in Raftlin knockdown cells (Fig. 2B, middle panels), consistent with the previous observation that knockdown of TICAM-2 or inhibition of clathrin-dependent endocytosis reduced the LPS-induced IL-6 production in RAW 264.7 cells (26). In contrast, LPS-induced TICAM-1–independent cytokine expression such as TNF-α was almost unaffected by knockdown of Raftlin in Mo-Mϕs (Supplemental Fig. 1A). Thus, Raftlin regulates endosomal TLR4 signaling in human primary cells. Notably, Raftlin expression was upregulated by LPS stimulation in Mo-Mϕs (Supplemental Fig. 1B).

The internalization of LPS-bound TLR4 dimer is a prerequisite for TICAM-2–TICAM-1–mediated IFN-β production. Because the clathrin-mediated endocytic pathway is involved in TLR4 internalization in human epithelial cells (15), we examined the participation of the endocytic pathway in LPS-induced IFN-β production in human Mo-DCs and Mo-Mϕs using endocytic pathway inhibitors, chlorpromazine (clathrin-mediated endocytosis inhibitor), and methyl-β-cyclodextrin (lipid raft–dependent endocytosis inhibitor). LPS-induced IFN-β production was inhibited by pretreatment of cells with chlorpromazine, but not with methyl-β-cyclodextrin (Fig. 3A). To further confirm the clathrin-mediated TLR4 endocytosis in Mo-DCs and Mo-Mϕs, we pretreated cells with Dynasore, a dynamin inhibitor, and examined LPS-induced TICAM-1–dependent IFN-β production and MyD88-dependent TNF-α production. In Mo-DCs and Mo-Mϕs, LPS-induced IFN-β but not TNF-α production was inhibited with Dynasore, indicating the clathrin-dependent TLR4 endocytosis (Fig. 3B).

We knocked down Raftlin in Mo-Mϕs and analyzed the internalization of TLR4 into the early endosomes in response to LPS, by confocal microscopy (Fig. 3C). TLR4 resided on the plasma membrane and inside the cells in unstimulated control cells. Upon LPS stimulation, TLR4 accumulated on the cell surface around 5 min and then internalized within 30 min. The colocalization of TLR4 and an early endosome marker Rab5a was clearly observed at 15 min (Fig. 3C, upper left panels). In contrast, when Raftlin was knocked down, TLR4 accumulated at the cell surface but was not endocytosed, remaining on the plasma membrane during LPS stimulation and never colocalizing with Rab5a (Fig. 3C, upper right panels). Quantitative analysis showed that colocalization coefficient between TLR4 and Rab5a was significantly decreased in Raftlin knockdown Mo-Mϕs (Fig. 3C, lower right panel). These results indicate that Raftlin is required for TLR4 internalization into the endosomes upon LPS stimulation. Because Mo-Mϕs express TLR4 at low level on the cell surface, we could not assess LPS-induced TLR4 endocytosis by flow cytometry (Supplemental Fig. 2).

Raftlin localizes to lipid rafts in B and T cells, but it predominantly localizes in the cytoplasm in human epithelial cells (18–20). In human Mo-Mϕs prepared from different healthy donors, Raftlin mainly resided in the cytoplasm in either of the cells (Fig. 4). In response to LPS stimulation, Raftlin moved from the cytoplasm to the plasma membrane and colocalized with TLR4 at the cell surface at 5 min and then internalized with TLR4 (Fig. 4). After 15 min, colocalization of Raftlin and TLR4 disappeared in Mo-Mϕs. The velocity of TLR4 endocytosis appears to depend on individual cells derived from different donors (Fig. 4, right panels).

The previous data showing that Raftlin mediates poly(I:C) cellular uptake through association with clathrin raises the possibility that Raftlin recognizes TLR4 as cargo in clathrin-dependent endocytosis upon LPS stimulation. To test this possibility, we assessed interaction of Raftlin with TLR4 and clathrin by coimmunoprecipitation assay using HEK293/TLR4/MD-2/CD14. Raftlin physically interacted with TLR4 and clathrin 10 min after LPS stimulation and then dissociated from TLR4 and clathrin after 30 min (Fig. 5A). Raftlin associated with clathrin adaptor AP-2 in HEK293-TLR4/MD2/CD14 cells independent of LPS stimulation (Fig. 5B). As previously reported (18), Raftlin transiently associated with clathrin in response to poly(I:C) stimulation in HeLa cells (Fig. 5C). Proteome analysis of Raftlin-binding proteins upon poly(I:C) stimulation in HeLa cells revealed that clathrin H chain was identified as a poly(I:C)-dependent Raftlin binding protein and that AP-2 was identified as Raftlin-binding protein at high levels in unstimulated cells (Fig. 5C). These results suggest that Raftlin interacts with AP-2 at steady-state and participates in the selection of cargo together with AP-2 after recruitment to the plasma membrane.

Mouse splenic DCs and BM-Mϕ express Raftlin and Raftlin2 in contrast with human DCs and macrophages that express only Raftlin (Fig. 6A). In mouse B cells, Raftlin2 compensates Raftlin function in Rftn^−/− cells (20). To determine the physiological function of Raftlin and Raftlin2, we generated Raftlin2-deficient (Rftn2^−/−) mice (Supplemental Fig. 3). Raftlin2-deficient offspring were born at the expected Mendelian ratio from intercrosses of heterozygotes (Supplemental Fig. 3). The complete absence of Raftlin2 mRNA expression was confirmed by PCR (Fig. 6B). We further generated Raftlin and Raftlin2 double-knockout (Rftn^−/−/Rftn2^−/−) mice by Rftn2^−/− mice being crossed with Rftn^−/− mice (Supplemental Fig. 3). The knockout of Raftlin and/or Raftlin2 did not affect the expression of TLR3, TLR4, and TICAM-1 (Fig. 6B). These mice appeared healthy and showed normal development under specific pathogen-free conditions.

We analyzed cytokine production by splenic DCs, BM-Mϕs, and peritoneal Mϕs from wild-type, Rftn^−/−, Rftn2^−/−, or Rftn^−/−/Rftn2^−/− mice in response to poly(I:C) or LPS (Fig. 6C–E). Surface expression levels of TLR4 and CD14 in Rftn^−/−/Rftn2^−/− DCs and Mϕs were comparable with that of wild-type mice (Fig. 7A). Upon LPS stimulation, TICAM-1–dependent IFN-β production was partially decreased in Rftn^−/− or Rftn2^−/− splenic DCs and greatly reduced in Rftn^−/−/Rftn2^−/− DCs (Fig. 6C). TNF-α production was unaffected by Raftlin and Raftlin2 double knockout, whereas IL-6 production was partially decreased in Rftn^−/−/Rftn2^−/− splenic DCs compared with wild-type DCs (Fig. 6C). Poly(I:C) induced IFN-β, TNF-α, and IL-6 production depending on Raftlin/Raftlin2. Notably, knockout of Raftlin and/or Raftlin2 minimally affected Pam₂CSK4 (TLR2/6 ligand) response (Fig. 6C). Thus, Raftlin and Raftlin2 participate in the endocytosis of TLR3/4 ligands in mouse splenic DCs. In contrast, LPS-induced IFN-β and IL-6 production was slightly decreased in Rftn^−/−/Rftn2^−/− BM-Mϕs, whereas TNF-α production was unaffected by knockout of Raftlin and Raftlin2 (Fig. 6D). In peritoneal macrophages from Rftn^−/−/Rftn2^−/− mice, the levels of cytokines induced by poly(I:C) or LPS were comparable with those in macrophages from wild-type mice (Fig. 6E). IFN-β was scarcely produced in peritoneal Mϕs irrespective of genotype in response to LPS.

We then analyzed LPS-induced TLR4 endocytosis in BM-Mϕs from wild-type and Rftn^−/−/Rftn2^−/− mice by flow cytometry. Consistent with the recently reported data (16, 27), LPS induced internalization of TLR4 and CD14 in wild-type BM-Mϕs (Fig. 7B). Double knockout of Raftlin and Raftlin2 slightly affected LPS-induced TLR4 endocytosis in BM-Mϕs (Fig. 7B). Thus, in BM-Mϕs, Raftlin/Raftlin2 appear to be marginally involved in LPS-induced TLR4 endocytosis. It has been reported that CD14 controls the LPS-induced TLR4 endocytosis in BM-Mϕs (16, 27). The difference in the dependency of Raftlin/Raftlin2 in LPS-induced TLR4 endocytosis between splenic DCs and BM-Mϕs may result from the expression level of surface CD14. Indeed, CD14 level of BM-Mϕs was higher than that of splenic DCs (Fig. 7A). These results indicate that requirement of Raftlin and Raftlin2 in LPS-induced TLR4 endocytosis and TICAM-1–dependent IFN-β production appears to depend on cell types in mice.

The endocytosis of microbial components is essential for endosomal TLRs to induce innate immune responses. In contrast with endosome/lysosome-localizing TLR3, 7, 8, and 9, cell surface–localizing TLR4 must be internalized upon LPS binding to activate the TICAM-2–TICAM-1 pathway at endosomes. The clathrin- and dynamin-mediated endocytic pathway is involved in TLR4 internalization in human epithelial cells (15). Also, membrane-bound CD14 has been shown to mediate LPS-induced TLR4 internalization in mouse DCs and Mϕs (16, 27), but the relationship between CD14-mediated and clathrin/dynamin-mediated TLR4 endocytosis remains unknown. Recently, Klein et al. (28) reported that LPS stimulation induced clustering of TLR4 into small punctate structures in the plasma membrane that contain CD14/LPS and clathrin in HEK293/TLR4/MD2/CD14 cells and CD14⁺ macrophage cell lines. In this study, we identified Raftlin as a new regulator of TLR4 endocytosis in human Mo-DCs and Mo-Mϕs, which is involved in clathrin-mediated endocytosis. Because surface CD14 level of Mo-DCs and Mo-Mϕs is negative and low, respectively (29–31) (Supplemental Fig. 1), Raftlin-dependent TLR4 endocytosis in Mo-DCs and Mo-Mϕs appears to be independent of CD14-mediated signaling. Indeed, Mortellaro and colleagues (31) demonstrated that LPS stimulation did not induce Syk phosphorylation downstream of CD14 in Mo-DCs and Mo-Mϕs. In contrast, CD14 was required for LPS-induced TLR4 endocytosis in human monocytes that express CD14 at high level (31), suggesting that surface CD14 level might affect the dependency of Raftlin in TLR4 endocytosis upon LPS simulation.

Localization and signaling of TLR4 is regulated by many proteins/mechanisms before and after LPS recognition. RAB11A GTPase is involved in the recruitment of TLR4 from the endocytic recycling compartment to E.coli phagosome (32) and affects localization of TICAM-2 to the endocytic recycling compartment and early sorting endosomes (32). The small GTPase ADP-ribosylation factor 6 also regulates transport of the TLR4 adaptor TIRAP and TICAM-2 to distinct signaling platforms (33). In addition, multiple molecules such as CD14, MD-2, PI3K, and Raftlin regulate TLR4 internalization (16, 17, 31), which is important in shifting the TLR4 signaling axis from MyD88 to TICAM-1, leading to expression of IFN-β and IFN-stimulated genes including caspase 4/11, a cytoplasmic LPS receptor (34, 35). Hence LPS-induced dynamic redistribution and intracellular trafficking of TLR4 and its adaptor molecules control innate immune signaling against microbial infection.

In clathrin-dependent endocytosis, AP-2 interacts with PI(4,5)P₂ via the C-terminal domain of μ subunit after recruitment to the plasma membrane, followed by conformational change and recognition of cargo protein, which triggers clathrin assembly (36). Many accessary proteins interacting with the α or β appendage of AP-2 have been reported to modulate cargo selection and clathrin-coated vesicle formation (37). Considering that Raftlin colocalized with TLR4 upon LPS stimulation in Mo-Mϕs (Fig. 4), Raftlin might participate in cargo selection as an accessary molecule of clathrin–AP-2 complex after recruitment to the plasma membrane in Mo-DCs and Mo-Mϕs. Notably, expression of Raftlin was upregulated by LPS stimulation in Mo-Mϕs and Mo-DCs (Supplemental Fig. 1 and data not shown), suggesting a positive feedback function of Raftlin in microbial infection. In addition, a recent study showed that LPS induced synthesis and secretion of Raftlin in HUVECs, and Raftlin levels in serum were higher in patients with septic shock than healthy control subjects, although detailed mechanism remains unanalyzed (38).

In B and T cells, Raftlin localizes to lipid rafts and modulates lipid raft–dependent BCR and TCR signaling (19, 20). In human Mo-DCs and epithelial cells, Raftlin predominantly localizes in the cytoplasm and controls clathrin-mediated endocytosis of microbial cargo such as dsRNA and virus-derived ssRNA harboring stem structures (18, 39). Thus, taken together with the present results, Raftlin functions in both innate and adaptive immunity in a cell type–dependent mode.

The synthetic TLR ligands have been applied to next-generation vaccine adjuvants for infectious diseases and tumors (40). We recently showed that synthetic TLR3-specific ligand is a pivotal vaccine adjuvant for antitumor immunotherapy (41). TICAM-1 signaling initiated from endosomal compartment plays key roles in activation of NK cells and CTLs in professional Ag-presenting DCs (41, 42). Efficient delivery of TLR3/4 ligands into early endosome of myeloid DCs is crucial for induction of antitumor and antimicrobial responses. The endocytic pathway is regulated by several molecules/mechanisms including accessory proteins for AP-2, Rab GTPases, and phosphatidylinositide conversion (43, 44). Further study on the molecular mechanisms of cargo selection and endosomal delivery provides new insight into the regulation of endosomal TLR-mediated signaling in innate and adaptive immunity.

We are grateful to the laboratory members for valuable discussions. We thank Dr. C. Obuse (Hokkaido University) for mass analysis; Drs. K. Miyake and T. Taniguchi (University of Tokyo), Dr. H. Nishimura (University of Tsukuba), and Drs. T. Fujita and S. Nagata (Kyoto University) for providing the plasmids; and Dr. A. Yoshimura for providing the Raftlin knockout mice. We also thank Akiko Morii-Sakai for technical assistance.

The authors have no financial conflicts of interest.