The first two authors contributed equally to this work.
The Toll-like receptor 7 (TLR7)-specific stimulus loxoribine uncovers a strong relationship within the TLR7, 8 and 9 subfamily
Version of Record online: 29 SEP 2003
Copyright © 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
European Journal of Immunology
Volume 33, Issue 11, pages 2987–2997, November 2003
How to Cite
Heil, F., Ahmad-Nejad, P., Hemmi, H., Hochrein, H., Ampenberger, F., Gellert, T., Dietrich, H., Lipford, G., Takeda, K., Akira, S., Wagner, H. and Bauer, S. (2003), The Toll-like receptor 7 (TLR7)-specific stimulus loxoribine uncovers a strong relationship within the TLR7, 8 and 9 subfamily. Eur. J. Immunol., 33: 2987–2997. doi: 10.1002/eji.200324238
- Issue online: 29 SEP 2003
- Version of Record online: 29 SEP 2003
- Manuscript Accepted: 29 AUG 2003
- Manuscript Revised: 1 AUG 2003
- Manuscript Received: 4 JUN 2003
- Cellular activation;
- Molecular biology;
- Toll-like receptor;
- endosomal maturation
Loxoribine (7-allyl-7,8-dihydro-8-oxo-guanosine) acts as synthetic adjuvant in anti-tumor responses. Here we first demonstrate that loxoribine activates cells of the innate immune system selectively via the Toll-like receptor (TLR) 7/MyD88-dependent signaling pathway. TLR7- and MyD88-deficient immune cells fail to proliferate or produce cytokines in response to loxoribine, and genetic complementation of TLR7-deficient cells with murine or human TLR7 confers responsiveness. Subsequently we show that cellular activation by loxoribine and resiquimod (R-848), a stimulus for TLR7 and TLR8, depends on acidification and maturation of endosomes and targets MyD88 to vesicular structures with lysosomal characteristics. This mode of TLR7 and TLR8 action resembles CpG-DNA-driven TLR9 activation. We thus conclude that TLR7, 8 and 9 form a functional subgroup within the TLR family that recognizes pathogen-associated molecular patterns in endosomal/lysosomal compartments.
Extracellular signal-regulated kinase
Green fluorescent protein
c-Jun N-terminal kinase
Lysosomal-associated membrane protein-1
Myeloid differentiation primary response gene 88
Pathogen-associated molecular pattern
The innate immune system has developed pattern recognition receptors (PRR) that promote rapid responses to microbial pathogens. Recognition of conserved pathogen-associated molecular patterns (PAMP) leads to the activation of innate immune cells directing the emanating adaptive response 1. Toll-like receptors (TLR) sense PAMP and activate immune cells via the MyD88-dependent TLR/IL-1R signaling pathway which involves downstream mediators such as myeloid differentiation primary response gene 88 (MyD88), IL-1 receptor-associated kinase (IRAK) and TNF receptor-associated factor 6 (TRAF-6). Oligomerization of TRAF-6 leads to the activation of mitogen-activated protein kinases (MAPK), the stress-activated protein kinase c-Jun N-terminal kinase (JNK), p38 and thetranscription factors activating protein-1 (AP-1) and nuclear factor-κB (NF-κB) 2, 3.
Ten members of the TLR family (TLR1–10) have been reported in human: TLR2 is essential for the recognition of peptidoglycan, lipopeptides, heat shock proteins or lipoteichoic acid, while TLR4 is activated by lipopolysaccharide and heat shock proteins 4–7. TLR3 mediates the recognition of double-stranded viral RNA 8, TLR5 activates immune cells in response to flagellin 9 and TLR9 recognizes unmethylated bacterial CpG-DNA 10, 11. We recently identified a synthetic antiviral compound, the imidazoquinoline resiquimod (R-848), as an activator of human and murine TLR7 12 and human, but not murine, TLR8 13.
According to the current view, microbial components act as natural adjuvants for the adaptive immune response, leading to the expression of costimulatory molecules and the secretion of proinflammatory and regulatory cytokines 14. Synthetic low-molecular-weight immune modifiers such as substituted guanosine nucleosides and imidazoquinolines are prominent examples of synthetic adjuvants that induce cytokine patterns similar to those of natural microbial TLR-stimulating ligands 15, 16. Substituted low-molecular-weight nucleosides, such as loxoribine (7-allyl-7,8-dihydro-8-oxo-guanosine), 8-bromo-guanosine or 8-mercapto-guanosine activate immune cells and lead to cytokine secretion 17, 18. Loxoribine, the more potent guanosine analog, shows anti-viral and anti-tumor activity in murine animal models and has entered phase I clinical trials 19–21. Based on their effective anti-viral and anti-tumor activities in various animal models, the imidazoquinolines imiquimod and resiquimod (R-848) have been approved for treatment of genital warts caused by human papillomavirus, and the more potent compound R-848 is anticipated as a treatment of recurring genital herpes 22–25. The finding that human and murine TLR7 12 and human TLR8 13 confer responsiveness to R-848 supports the idea that synthetic adjuvants may signal via TLR and thus mimic their as-yet-unknown natural ligands.
Here we show that the adjuvant loxoribine (7-allyl-7,8-dihydro-8-oxo-guanosine) activates the MyD88-dependent TLR/IL-1R signaling pathway specifically via TLR7, since proliferation of splenocytes and cytokine production by macrophages is abolished in TLR7- and MyD88-deficient mice. Activation of TLR7 is sensitive to bafilomycin A1 and can be blocked by inhibiting early endosome fusion via dominant negative Rab5 GTPase. Furthermore, stimulation of macrophages with TLR7 ligands results in recruitment of MyD88 to cellular compartments that stain positive for the lysosomal marker protein LAMP-1. Human TLR8 activation via the imidazoquinoline resiquimod (R-848) is also dependent on endosomal maturation (acidification), since bafilomycin A1 blocks activation. Overall these data imply that, similar to bacterial CpG-DNA-induced TLR9 activation, loxoribine- and imidazoquinolin-driven signaling via TLR7 and 8 require endosomal maturation and are dependent on MyD88 recruitment for initiation of downstream signaling events.
2.1 TLR7 confers responsiveness to loxoribine
In the quest to identify specific stimuli for TLR7, we screened low molecular weight immune modifiers for their potential to activate HEK 293 cells transiently transfected with human TLR2, 3, 4, 7, 8 or 9 cDNA and a NF-κB luciferase reporter plasmid. As shown in Fig. 1A, the guanosine analog loxoribine specifically signaled via TLR7, as human TLR2, 3, 4, 8 and 9, which responded to their cognate PAMP (Fig. 1B), did not respond to loxoribine. In this regard loxoribine differs from the imidazoquinoline R-848, which has recently been shown to activate human TLR7 and TLR8 (13 and Fig. 1B). Loxoribine and R-848 dose-dependently activated NF-κB in HEK 293 cells transfected with either human or murine TLR7 (Fig. 1C, D). Interestingly, the guanosine analogs carrying a different (non-oxygen) atom at position 8 (8-bromo-guanosine, 8-mercapto-guanosine) or analogs without the ribose moiety (the antiviral drugs gancyclovir and acyclovir) were unable to activate NF-κB (Fig. 1E, F). Collectively, these data demonstrate that loxoribine specifically signals via human and murine TLR7.
2.2 Responses to loxoribine in TLR7-deficient primary immune cells
Since loxoribine is mitogenic for B cells 26 and triggers cytokine production in macrophages 17, 27, we analyzed these parameters in cells from TLR7-deficient mice 12. While spleen cells from wild-type mice responded in a dose-dependent manner to loxoribine, splenocytes from TLR7-deficient mice failed to proliferate (Fig. 2A). Furthermore, peritoneal macrophages from wild-type mice produced TNF-α and IL-12p40 in a dose-dependent fashion upon loxoribine stimulation, while TLR7-deficient macrophages did not respond (Fig. 2B). Overall, these data prove an essential role for TLR7 in loxoribine-mediated signaling.
2.3 Loxoribine-induced TLR7 signaling via p38 and ERK is MyD88-dependent
The adaptor molecule MyD88 is essential for signaling initiated by TLR2, 3, 5 and 9 2, 14, 28, 29. To address whether loxoribine-mediated TLR7 signaling is MyD88-dependent, we stimulated spleen cells from wild-type and MyD88-deficient mice with loxoribine. As shown in Fig. 3A, splenocytes from wild-type mice proliferated vigorously in response to loxoribine, while spleen cells from MyD88-deficient mice responded only to ionomycin/TPA. Furthermore, spleen cells from wild-type mice, but not cells from MyD88-deficient mice, produced IL-6, IL-12p40 and TNF-α upon loxoribine stimulation (Fig. 3B–D). In addition, loxoribine-driven, but not TNF-α–driven, NF-κB activation was dose-dependently blocked in HEK293 cells transiently transfected with human TLR7 and increasing amounts of dominant negative human MyD88 (MyD88-C) (Fig. 3E). These data reveal an essential role for MyD88 in loxoribine-driven TLR7 signaling. Finally, we analyzed the activation of ERK and p38 in response to loxoribine (Fig. 3F). In wild-type macrophages, loxoribine induced ERK and p38 phosphorylation in a time-dependent manner, whereas these signaling events were completely abolished in macrophages from TLR7-deficient mice. In contrast, LPS stimulation activated ERK and p38 in wild-type and TLR7-deficient macrophages, demonstrating the dependency of loxoribine-mediated signaling on TLR7.
2.4 The TLR7 stimuli loxoribine and R-848 lead to recruitment of MyD88 to vesicular structures
Since loxoribine-induced signaling via TLR7 is MyD88-dependent, we visualized recruitment of MyD88 to display the cellular compartment where signaling is initiated. We also included the TLR7 ligand R-848 in these studies 12, 13. RAW264.7 cells stably transfected with a green fluorescent protein (GFP)-MyD88 fusion protein 30 were stimulated with loxoribine or R-848, and localization of the fusion protein was assessed by confocal microscopy (Fig. 4A). Loxoribine-driven MyD88 recruitment to intracellular vesicles is first visible 90–120 min after stimulation. R-848 induced similar recruitment, which could be observed within 30 min. Based on these observations, we analyzed in parallel the time-dependent phosphorylation of ERK, p38 and JNK and the degradation of IκBα in the murine macrophage cell line RAW264.7 (Fig. 4B). Loxoribine-induced phosphorylation of ERK, p38 and JNK was detectable within 10 min, reached its peak at 40 min and persisted up to 90 min (Fig. 4B). Degradation of IκBα was observed between 20 and 40 min after stimulation. In contrast, R-848-induced phosphorylation of the same kinases was detected within 5 min and peaked at 10–20 min (Fig. 4B). Degradation of IκBα could be observed within 5 min.
We conclude that the signaling induced by both loxoribine and R-848 is initiated at intracellular vesicles after MyD88 recruitment and subsequently leads to phosphorylation of kinases and IκBα degradation. The visualization of MyD88 recruitment is delayed due to the lower sensitivity of confocal microscopy as compared to Western blotting. Interestingly, the TLR7 stimulus loxoribine activates immune cells notably slower than R-848, suggesting affinity differences between these compounds or involvement of additional receptors.
2.5 TLR7 signaling requires endosomal maturation
Among the TLR characterized so far, only CpG-DNA-driven signaling via TLR9 requires acidification and therefore maturation of endosomes 30, 31. CpG-DNA signaling is efficiently blocked by a dominant negative version of the Rab5 GTPase, which interferes with early steps in endosomal fusion, as well as by bafilomycin A1, a specific inhibitor of the V-type ATPase responsible for acidification of endosomes and lysosomes 30–33.
To assess TLR7 dependency on endosomal maturation, we analyzed the effect of the dominant negative Rab5 GTPase (S34N) on loxoribine- and R-848-mediated signaling in HEK 293 cells transiently transfected with murine TLR7, a NF-κB luciferase reporter plasmid and increasing amounts of a dominant negative Rab5GTPase (S34N) cDNA expression construct. Mutated Rab5GTPase inhibited loxoribine- and R-848-driven NF-κB activation in a dose-dependent fashion but did not affect signal transduction induced by TNF-α (Fig. 5A). We used the murine macrophage cell line RAW264.7 to examine the effect of bafilomycin A1 on phosphorylation of downstream kinases upon TLR7 stimulation 34, 35. Like CpG-ODN-induced phosphorylation via TLR9, loxoribine- and R-848-driven phosphorylation of ERK was efficiently blocked by bafilomycin A1 (Fig. 5B). In contrast, TLR2- and TLR4-driven phosphorylation of ERK by Pam3Cys and LPS, respectively, was insensitive to bafilomycin A1. TNF-α production by Raw264.7 cells followed the same pattern seen with ERK phosphorylation (data not shown).
Bafilomycin A1 also blocked NF–κB activation induced by loxoribine, R-848 and CpG-ODN in a dose-dependent fashion in HEK 293 cells transiently transfected with TLR7 (Fig. 5C) or TLR9 (Fig. 5D). However, TNF-α–driven NF–κB activation in HEK 293 cells was insensitive to bafilomycin A1. TLR8, known to be highly homologous to TLR7 and TLR9 36, 37, shares the same functional dependency on endosomal maturation since bafilomycin A1 blocked R-848-induced NF–κB activation in HEK 293 cells transiently transfected with human TLR8 (Fig. 5E).
Bafilomycin A1 treatment also impaired loxoribine- and R-848-driven MyD88 recruitment to intracellular vesicles (Fig. 6). RAW264.7 cells expressing GFP-MyD88 fusion protein were preincubated with bafilomycin A1 and stimulated with loxoribine, R-848 or Pam3Cys. Bafilomycin completely inhibited loxoribine- or R-848-induced MyD88 recruitment but did not influence MyD88 recruitment upon Pam3Cys stimulation. Of note, TLR2-mediated signaling recruited GFP-MyD88 fusion protein predominantly to the inner surface of the plasma membrane and not into vesicular structures as seen with TLR7 stimulation. These observations suggest that MyD88 recruitment to intracellular vesicles is an essential part of the signaling events initiated by TLR7.
Collectively, these data demonstrate that stimulation of TLR7 by loxoribine and R-848, as well as R-848-driven signaling via TLR8, depends on endosomal maturation and thus resembles CpG-DNA-driven signaling via TLR9.
2.6 TLR7 stimulation recruits MyD88 to endosomal/lysosomal compartments
Activation through TLR7 requires endosomal maturation and leads to recruitment of MyD88 to cytoplasmic vesicular structures (see above). CpG-ODN-driven stimulation via TLR9 also targets MyD88 to vesicular structures; these vesicles contain TLR9 and stain for the endosomal/lysosomal marker LAMP-1 (lysosomal-associated membrane protein-1, CD107A) 30. To address the question of whether TLR7-initiated signaling recruits MyD88 to endosomal/lysosomal structures, we examined localization of the GFP-MyD88 fusion protein in RAW264.7 macrophages after stimulation with TLR7 ligands. Confocal microscopy demonstrated that upon stimulation with loxoribine or R-848, a predominant colocalization (yellow, indicated by arrows) of GFP-MyD88 (green fluorescence) and LAMP-1 (red fluorescence) occurred (Fig. 7). In contrast, Pam3Cys-induced TLR2 signaling did not lead to colocalization of GFP-MyD88 and endosomal/lysosomal structures.
Since ligand-driven TLR7 activation recruits MyD88 to LAMP-1-positive vesicular structures (lysosomes), we suggest that TLR7 initiates signaling at the endosomal/lysosomal compartment, as shown previously for TLR9 30.
Toll-like receptors (TLR) of innate immune cells sense PAMP and lead to cell activation via the MyD88-dependent TLR/IL-1R signaling pathway 14, 38. We recently demonstrated that the imidazoquinoline resiquimod (R-848) activates immune cells via human and murine TLR7 12, as well as human TLR8 13. Of note, murine TLR8 does not confer responsiveness to R-848 and may be non-functional due to amino acid deletions in the extracellular domain 13.
Here we show that the synthetic low molecular weight immune modifier loxoribine (7-allyl-7,8-dihydro-8-oxo-guanosine) selectively activates cells via TLR7, differing from R-848, which triggers TLR7 and TLR8 (Fig. 1A, B). While loxoribine strongly activates NF-κB in murine and human TLR7-transfected HEK 293 cells, other modified guanosines, such as 8-bromo-guanosine and 8-mercapto-guanosine, and the antiviral drugs acyclovir and gancyclovir failed to do so (Fig. 1E, F). Loxoribine leads to the secretion of inflammatory and regulatory cytokines such as IL-6, TNF-α, and IL-12 in a MyD88-dependent manner (Fig. 3A–E). Cells from TLR7-deficient mice failed to respond to loxoribine by proliferation or cytokine production, proving that TLR7 is essential for recognition of the adjuvant loxoribine (Fig. 2).
In terms of the signal pathways involved, stimulation with loxoribine activates and phosphorylates JNK, ERK and p38. This activation is TLR7-dependent, as TLR7-deficient cells fail to induce phosphorylation of the kinases in response to loxoribine (Fig. 3F). Triggering by loxoribine and R-848 resulted in different dynamics of murine TLR7-driven JNK, ERK and p38 activation; loxoribine-induced phosphorylation is weaker and slower than that driven by R-848 (Fig. 4B). The difference in activation strength and time kinetics (Fig. 4A–B) is reflected in the concentrations of loxoribine and R-848 needed for cell stimulation. A 500-fold higher (2 μM versus 1 mM) concentration of loxoribine is necessary to obtain a response comparable to R-848 (Fig. 1) 12. The chemical structure of the two synthetic compounds is quite different, but they share a purine base: Loxoribine is a guanosine analog, whereas R-848 can be considered as an analog of adenine (Fig. 8). This common purine content may partly account for the recognition via TLR7, whereas the structural differences further affect the affinity, uptake or stability of the compounds. Indeed, the requirement of high loxoribine concentrations for efficient stimulation could reflect low-affinity binding to the TLR7 signaling receptor. Alternatively, inefficient cellular uptake of this guanosine analog into the cell (see below) could be a bottleneck for cellular activation. Since immobilized nucleosides are biologically inactive, it can be concluded that cellular uptake is a prerequisite for cell activation 39. Guanosine nucleosides, in contrast to R-848, may also be metabolized inside the cell, thus reducing the effective ligand concentration. On the other hand, the structure of R-848 is more hydrophobic; R-848 is likely to cross the membrane by diffusion rather than by receptor-mediated endocytosis (Fig. 8). Furthermore, R-848 could have a higher affinity towards the signaling receptor TLR7. Since R-848 does not activate murine TLR8 13, the rapid cellular activation of murine macrophages cannot be explained by simultaneous activation of murine TLR7 and TLR8. Chemical modifications of loxoribine may allow the generation of new synthetic analogs that specifically trigger TLR7 at much lower concentrations and could lead to a new class of anti-viral drugs or adjuvants for tumor vaccines.
TLR1, 2, and 4 are located on the cell surface and get recruited to phagosomes upon ligand-driven TLR activation 40–42. TLR9 is the only TLR family member known to be located in endosomal/lysosomal compartments but not at the cell surface 30. Furthermore, TLR9 signaling requires CpG-ODN uptake and endosomal maturation, which can be blocked by bafilomycin A1 or dominant negative Rab5 GTPase 31. Upon TLR9 activation MyD88 is recruited to LAMP-1-positive 43 endosomal/ lysosomal vesicles where signaling is initiated 30. Based on sequence homology, TLR7, TLR8 and TLR9 are believed to form an evolutionarily conserved cluster within the TLR family 36, 37. To analyze whether the sequence homology of TLR7, 8 and 9 translates into functional homology, we searched for hallmarks of TLR9 signaling, such as endosomal maturation or recruitment of MyD88 to lysosomal compartments, in TLR7- and TLR8-driven cell activation. Indeed, stimulation with loxoribine and R-848 resulted in recruitment of MyD88 to intracellular LAMP-1-positive vesicles (Fig. 7), and JNK, p38, ERK and NF-κB activation as well as TNF-α secretion were found to be sensitive to bafilomycin A1 or dominant negative Rab5 GTPase (Fig. 5, 6 and data not shown). In contrast, TLR2 and TLR4 signaling via their respective ligands, Pam3Cys and LPS, resulted in recruitment of MyD88 to the cell membrane and was independent of endosomal maturation (Fig. 6, 7) 39. Since R-848-driven NF-κB activation in human TLR8-transfected HEK 293 cells is also blocked by bafilomycin A1 (Fig. 5E), the TLR family members segregate functionally into two groups. TLR7, 8 and 9 share a functional dependency on endosomal maturation for initiation of cellular activation and are thus specialized to sense PAMP in endosomal/lysosomal compartments. TLR2 and TLR4, in contrast, recognize their natural PAMP at the cell surface and initiate signaling at the cell membrane.
Overall our results imply that human and murine TLR7, as well as human TLR8, mediate recognition of their as-yet-unknown natural ligands within the endosomal/lysosomal compartment. Thus, we may deduce that pathogen-derived TLR7 and 8 ligands must be liberated by endosomal/lysosomal processing. The synthetic adjuvant loxoribine presumably mimics the natural PAMP for TLR7, leading us to believe that RNA breakdown products containing modified guanosines are likely candidates for natural TLR7 ligands.
4 Materials and methods
4.1 Cells and reagents
Human embryonic kidney 293 T cells and the murine macrophage cell line RAW267.4 were obtained from ATCC and cultivated in Dulbecco's modified Eagle's medium (PAN, Aidenbach, Germany)or VLE-RPMI medium (Biochrom KG, Berlin, Germany) supplemented with 10% FCS, 100 IU/ml penicillin G and 100 IU/ml streptomycin sulfate. R-848 was commercially synthesized by GLSynthesis Inc. (Worcester, MA, USA) and provided by Coley Pharmaceuticals GmbH (Langenfeld, Germany). E. coli LPS 055:B5, poly I-C and loxoribine were from Sigma-Aldrich (Taufkirchen, Germany). CpG-ODN 1668 (TCCATGACGTTCCTGATGCT) was synthesized by TIB BIOMOL (Berlin, Germany) in a phosphothioate-protected form. Primers for TLR cDNA amplification were synthesized by MWG Biotech (Ebersberg, Germany). Bafilomycin A1, TNF-α and Pam3Cys were from Calbiochem (Bad Soden, Germany) and PeproTech Inc. (Rocky Hill, NJ, USA). Alexa Flour® 546-labeled goat anti-rat was from Molecular Probes (Leiden, The Netherlands). Antibody against LAMP-1 (CD107a, Clone 1D4B) was purchased from Becton Dickinson (Franklin Lakes, NJ, USA). Antibodies against phospho-p38 (Thr180/Tyr182), IκBα and phospho-ERK1/2 (Thr202/Tyr204) were obtained from New England Biolabs (Beverly, MA, USA). Anti-ERK-1, 2 and anti-p38 antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA).
4.2 Plasmids and cDNA
Expression plasmids for human TLR3 and TLR4, dominant negative human MyD88 (hMyD88-C) and human MD2 (hMD2) were kindly provided by Tularik Inc. (South San Francisco, USA) and Kensuke Miyake (Saga Medical School, Japan). Expression plasmids for human and murine TLR7 and human TLR8 and TLR9 have been described previously 11, 13. The Rab5 S34N mutant and generation of pEF-eGFP-MyD88 and eGFP-MyD88-expressing RAW264.7 macrophages were previously reported 30.
4.4 Transfection and reporter assays
For monitoring of transient NF-κB activation, 3×106 293 HEK cells were electroporated at 200 V and 960 μF with 100 ng (human TLR2, 4, MD-2) or 1 μg (human TLR3, 7, 8, 9 and murine TLR7) TLR expression plasmid and 20 ng NF-κB luciferase reporter plasmid (kindly provided by Patrick Baeuerle, Munich, Germany). For testing the dominant negative effect of hMyD88-C or Rab5 GTPase, 293 cells were transiently cotransfected with hTLR7, increasing amounts of MyD88-C or pCX-Rab5-S34N and 20 ng of sixfold NF-κB luciferase reporter plasmid. The overall amount of plasmid DNA was held constant at 15 μg per electroporation by addition of the appropriate empty expression vector. Cells were seeded at 105 cells/well and, after overnight culture, stimulated with 1 mM (or as indicated) loxoribine, 1 μM CpG-ODN 2006, 3 μM R-848, 100 ng/ml LPS, 100 μg/ml poly I-C or 5 μg/ml Pam3Cys for a further 7–10 h. In some experiments cells were incubated 15 min with bafilomycin A1 prior to stimulation. Stimulated cells were lysed using reporter lysis buffer (Promega, Mannheim, Germany), and lysate was assayed for luciferase activity using a Berthold luminometer (Wildbad, Germany) according to the manufacturer's instructions. Thioglycollate-elicited cells (TLR7-deficient) or spleen cells (MyD88-deficient) were seeded into 96-wellplates at a concentration of 5×104 cells/well or 1×106 cells/well and stimulated with the indicated concentrations of loxoribine or 1 μg/ml LPS in the presence or absence of 30 U/ml IFN-γ. Concentrations of TNF-α, IL-12 p40 or IL-6 in the culture supernatants were determined by ELISA according to the manufacturer's instructions (Genzyme Techne, Minneapolis, MN and BD PharMingen, San Diego, CA). For monitoring cell proliferation, 105 spleen cells were seeded in 96-well plates in 200 μl medium. After stimulation for 48 h, 20 μl [3H]-thymidine (ICN, Eschwege, Germany) was added, and cells were incubated for an additional 16 h before harvest.
4.5 Western blotting and immunocomplex kinase assays
Cells were stimulated with 3 μM R-848, 1 mM Loxoribine, 100 ng/ml LPS, 1 μM CpG-ODN 1668 or 5 μg/ml Pam3Cys. After stimulation, cells were lysed in Triton lysis buffer containing 25 mM Hepes (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 10% glycerol, 1% Triton X-100, 10 mM pyrophosphate, 20 mM β-glycerophosphate, 2 mM orthovanadate, 10 mM sodium fluoride, 10 μg/ml leupeptin and 1 mM phenylmethylsulfonyl fluoride (PMSF). Lysates were boiled in SDS sample buffer, sonicated, centrifuged at 10,000×g for 10 min, resolved by 10% SDS-PAGE and blotted onto Protran nitrocellulose membranes (Schleicher & Schuell, Dassel, Germany). Membranes were blocked in 5% skim milk solution, probed with the indicated antibodies and visualized using the Renaissance chemiluminescence kit (NEN). The activity of JNK kinase was assayed as previously described 30.
4.6 Confocal laser scanning microscopy of living and fixed cells
One day before image collection, 7×105 GFP-MyD88-expressing RAW264.7 macrophages were plated on round (diameter 12 mm) glass slides (Joseph Peske oHG, Aindling-Pichl, Germany). Cells were incubated with 5 μg/ml Pam3Cys, 1 mM loxoribine or 3 μM R848 for the indicated time. Stimulation was stopped by washing the cells in ice-cold PBS. Cells were subsequently fixed in 3% formalin for 15 min at room temperature, followed by an additional PBS wash. For immunolabeling of the LAMP-1-positive lysosomal compartment, cells were incubated with primary antibody (anti-LAMP-1, 1:500) in freshly prepared permeabilization buffer (PBS, 0.2% BSA, 0.2% saponin) for 45 min, followed by another 45 min incubation at RT with the secondary antibody (Alexa Flour 546-labeled goatanti-rat, 1:500 in permeabilization buffer). After the final PBS washes, glass slides were fixed onto objective slides using Entelan® (Merck, Darmstadt, Germany), and the slides were covered by mounting fluid (Labsystems Oy, Helsinki, Finland). Cells were viewed with a Zeiss (Carl Zeiss, Jena, Germany) LSM 510 confocal microscope equipped with LSM 510 software version 2.02, an Ar/Kr laser (458 and 488 nm) and two He/Ne lasers (543 nm and 633 nm). The lens used was a Plan-Neofluar 40 1.3 oil lens. The confocal pinhole setting was usually chosen to show details of the fluorescent structures associated with the GFP-fusion protein, providing a <1 μm section of the cell at a resolution of 1024×1024 pixels. Images were exported as single-image files in the TIF format with the LSM5 Image Browser (version 2.8).
We would like to thank Martin Mempel and Verena Voelcker (Department of Dermatology and Allergy, Biederstein, Technical University, Munich, Germany) for providing human keratinocyte cDNA and Klaus Pfeffer and Gabriele Köllisch for critically reading the manuscript. This work was supported by BMBF grant 0311675/CpG Immunopharmaceuticals GmbH, SFB 456 and Deutsche Forschungsgemeinschaft grant BA 1618/2–1.