The TLRs 7, 8, and 9 stimulate innate immune responses upon recognizing pathogen nucleic acids. U-rich RNA sequences were recently discovered that stimulate human TLR7/8-mediated or murine TLR7-mediated immune effects. In this study we identified single-stranded RNA sequences containing defined sequence motifs that either preferentially activate human TLR8-mediated as opposed to TLR7- or TLR7/8-mediated immune responses. The identified TLR8 RNA motifs signal via TLR8 and fail to induce IFN-α from TLR7-expressing plasmacytoid dendritic cells but induce the secretion of Th1-like and proinflammatory cytokines from TLR8-expressing immune cells such as monocytes or myeloid dendritic cells. In contrast, RNA sequences containing the TLR7/8 motif signal via TLR7 and TLR8 and stimulate cytokine secretion from both TLR7- and TLR8-positive immunocytes. The TLR8-specific RNA sequences are able to trigger cytokine responses from human and bovine but not from mouse, rat, and porcine immune cells, suggesting that these species lack the capability to respond properly to TLR8 RNA ligands. In summary, we describe two classes of single-stranded TLR7/8 and TLR8 RNA agonists with diverse target cell and species specificities and immune response profiles.

Toll-like receptors play an essential role in pathogen recognition and innate immune responses in mammals. In humans, 10 different TLRs were described as recognizing a variety of pathogen-associated molecular patterns from bacteria, viruses, and fungi. TLR7, 8, and 9 belong to the same subfamily of Toll-like receptors based on their genomic structure, sequence similarity, and homology (1). TLR7 and 8 are activated by small antiviral compounds such as imidazoquinolines, guanine nucleoside analogues, whole RNA from viruses, certain synthetic single-stranded oligoribonucleotides (ORNs)3 or double-stranded small interfering RNA sequences (2, 3, 4, 5, 6, 7, 8). TLR9 recognizes nonmethylated, CpG-containing DNA of bacterial or viral origin or synthetic single-stranded CpG oligodeoxynucleotides (ODNs) (9, 10, 11).

Whereas most TLRs are expressed on the cell surface, TLR7, 8, and 9 are located in intracellular endolysosomal compartments. TLR7, 8, and 9 have a unique pattern of cell type-specific expression that is thought to be responsible for different pathogen response profiles (12). Hereby, human TLR7 and TLR8 together display a broader cell type-specific distribution than TLR9. TLR7 mRNA was mainly detected in plasmacytoid dendritic cells (pDCs) and B cells, and only low mRNA expression was observed in monocytes, macrophages, myeloid DCs (mDCs), and monocyte-derived dendritic cells (DCs). In contrast, TLR8 mRNA was found to be highly expressed in monocytes but also in mDCs, monocyte-derived DCs, and macrophages (13, 14, 15, 16, 17, 18).

Specific natural viral but also eukaryotic RNA sequences have been identified that are capable of stimulating human TLR7 and TLR8 when incorporated into single-stranded synthetic ORNs. The human TLR7 and TLR8 immune response appears to be stimulated strongest with U-rich sequences (4, 19, 20, 21), although recent work has suggested that the TLR7 immune response lacks specificity for defined sequence motifs (22). TLRs are type I integral membrane glycoproteins consisting of leucine-rich repeats (LRR) in the ectodomain (ECD), and cytoplasmatic signaling domains known as Toll IL-receptor (TIR) domains. The TIR and ECD domains are joined by a single trans membrane helix, and the ECDs containing the LRRs have been inferred to be responsible for ligand recognition (23, 24).

In this study, we searched for and identified RNA sequence motifs that selectively stimulate human TLR7 and TLR8 or only TLR8. The sequence-dependent effects were analyzed on primary human immune cells and recombinant TLR-expressing cells, demonstrating that specific sequences can induce TLR7 and TLR8 effects but others stimulate only TLR8-specific signaling and, therefore, lack pDC-mediated effects like IFN-α production and pDC maturation. TLR8-dependent RNA sequences were not able to trigger TLR-mediated effects in mouse or rat immune cells, demonstrating that the murine TLR8 not only lacks the capability to respond to TLR8 small molecule ligands (25, 26) but also to TLR8 RNA ligands. However, combining poly(dT) ODN and TLR8 RNA ligands resulted in a regaining of TLR8-dependent effects, suggesting that murine TLR8 is in principle capable of signaling upon encountering natural RNA ligands and may have another defect.

The TLR7 and TLR8 agonist resiquimod (R-848 or S-28463) was commercially synthesized by GLSynthesis. The CpG ODNs 2395 and 10103 were provided by Coley Pharmaceutical. ORNs were obtained from BioSpring. All oligonucleotides were controlled for identity and purity by Coley Pharmaceutical and had undetectable endotoxin levels (<0.1 endotoxin unit/ml) as measured by the Limulus assay (BioWhittaker). Polyriboinosinic:polyribocytidylic acid (poly(rI:rC) was obtained from Sigma-Aldrich. Oligonucleotides were suspended in sterile endotoxin-free Tris-EDTA (ODN) (Sigma-Aldrich) or in DNase- and RNase-free water (ORN) (Invitrogen Life Technologies) and stored and handled under aseptic conditions to prevent contamination. Sequences are listed in Table I.

Table I.

ORN sequencesa

ORN No.RNA Sequence
R-1075 CCGUCUGUUGUGUGACUC 
R-1263 GCCACCGAGCCGAAGGCACC 
R-0001 UAUAUAUAUAUAUAUAUAUA 
R-0002 UUAUUAUUAUUAUUAUUAUU 
R-0003 UUUAUUUAUUUAUUUAUUUA 
R-0004 UUUUAUUUUAUUUUAUUUUA 
R-0005 UGUGUGUGUGUGUGUGUGUG 
R-0006 UUGUUGUUGUUGUUGUUGUU 
R-0007 UUUGUUUGUUUGUUUGUUUG 
R-0008 UUUUGUUUUGUUUUGUUUUG 
Control ORN 1 CCGAGCCGCCGCCCCC 
Control ORN 2 CCGAGCCGAAGGCACC 
R-1211 dNdNdNdNdNGUUGUGUdNdNdNdNdNdN 
R-1212 dNdNdNdNdNdNUUGUGdNdNdNdNdNdNdN 
R-1213 dNdNdNdNdNdNUUGUdNdNdNdNdNdNdNdN 
R-1214 dNdNdNdNdNdNUUGdNdNdNdNdNdNdNdNdN 
R-1215 dNdNdNdNdNdNdNUGdNdNdNdNdNdNdNdNdN 
R-1216 dNdNdNdNdNdNdNdNGUdNdNdNdNdNdNdNdNdN 
R-1312 GCCACCGAGCGUUGUGUACC 
R-1321 GCCACCGAGCAUUGUGAACC 
R-1322 GCCACCGAGCAUUGAGAACC 
R-1323 GCCACCGAGCAAUGAGAACC 
R-1324 GCCACCGAGCAAGGUGAACC 
ORN No.RNA Sequence
R-1075 CCGUCUGUUGUGUGACUC 
R-1263 GCCACCGAGCCGAAGGCACC 
R-0001 UAUAUAUAUAUAUAUAUAUA 
R-0002 UUAUUAUUAUUAUUAUUAUU 
R-0003 UUUAUUUAUUUAUUUAUUUA 
R-0004 UUUUAUUUUAUUUUAUUUUA 
R-0005 UGUGUGUGUGUGUGUGUGUG 
R-0006 UUGUUGUUGUUGUUGUUGUU 
R-0007 UUUGUUUGUUUGUUUGUUUG 
R-0008 UUUUGUUUUGUUUUGUUUUG 
Control ORN 1 CCGAGCCGCCGCCCCC 
Control ORN 2 CCGAGCCGAAGGCACC 
R-1211 dNdNdNdNdNGUUGUGUdNdNdNdNdNdN 
R-1212 dNdNdNdNdNdNUUGUGdNdNdNdNdNdNdN 
R-1213 dNdNdNdNdNdNUUGUdNdNdNdNdNdNdNdN 
R-1214 dNdNdNdNdNdNUUGdNdNdNdNdNdNdNdNdN 
R-1215 dNdNdNdNdNdNdNUGdNdNdNdNdNdNdNdNdN 
R-1216 dNdNdNdNdNdNdNdNGUdNdNdNdNdNdNdNdNdN 
R-1312 GCCACCGAGCGUUGUGUACC 
R-1321 GCCACCGAGCAUUGUGAACC 
R-1322 GCCACCGAGCAUUGAGAACC 
R-1323 GCCACCGAGCAAUGAGAACC 
R-1324 GCCACCGAGCAAGGUGAACC 
a

All ORN were phosphorothioate modified (d refers to deoxyribonucleotides). Control ORN 1 and 2 were used for motif analysis, and the underlined nucleotides were exchanged for 4-mer sequence alterations as shown in Fig. 2.

Human embryonic kidney (HEK) cells (HEK293) containing a NF-κB-luciferase reporter construct and expressing human TLR7, TLR8, or TLR9 or without TLR expression were used as described before (25, 27). Cells were plated on 96-well plates at 1.5 × 104/well and allowed to attach overnight. The cells were subsequently incubated for 16 h with the indicated amount of ORNs complexed to 1,2-dioleoyloxy-3-(trimethylammonium)propane (DOTAP; Roche), DOTAP alone, R-848, poly(rI:rC), or CpG ODNs and then tested for luciferase expression. Each data point was done in duplicate.

HEK293T cells were plated on 96-well plates at 1.5 × 104/well and allowed to attach overnight 1 day before transfection. Transient transfection was performed with Effectene transfection reagent (Qiagen) according to the manufacturer’s protocol. To reach the appropriate level of expression of transfected genes, cells were incubated 24 h under their normal growth conditions. Stimulation of transfected cells was then performed with the indicated amounts of ORNs complexed to DOTAP (Roche), DOTAP alone, or R-848.

PBMC preparations from healthy male and female human donors were obtained from the Institute for Hemostaseology and Transfusion Medicine of the University of Düsseldorf (Düsseldorf, Germany). PBMCs were purified by centrifugation over Ficoll-Hypaque (Sigma-Aldrich). Purified PBMCs were washed twice with 1× PBS and resuspended in RPMI 1640 culture medium supplemented with 5% (v/v) heat-inactivated human AB serum (BioWhittaker) or 10% (v/v) heat inactivated FCS, 1.5 mM l-glutamine, 100U/ml penicillin, and 100 mg/ml streptomycin (all from Sigma-Aldrich).

Freshly isolated PBMCs were resuspended at a concentration of 3 × 106/ml to 5 × 106/ml and added to 96-well round-bottom plates (200 μl/well) that had previously received nothing or R-848, ORNs complexed to DOTAP, ODNs, or DOTAP alone. Cells were cultured in a humidified incubator at 37°C for the indicated time points. Culture supernatants (SN) were collected and, if not used immediately, frozen at −20°C until required. Amounts of cytokines in the SN were assessed using commercially available ELISA kits (IL-6, IL-10, IFN-γ, or TNF-α from Hölzel Diagnostika (Diaclone) or IL-12p40 from BD Pharmingen), in-house ELISAs (IFN-γ-inducible protein-10 (IP-10) or IFN-α) developed using commercially available Abs (from BD Pharmingen or PBL Biomedical Laboratories, respectively) or Luminex technology using a Cytokine 25-plex kit (BioSource International).

Human pDCs and monocytes were isolated from PBMCs by blood DC Ag (BDCA)-4 or CD14 immunomagnetic bead positive selection according to the manufacturer’s recommendations (Miltenyi Biotech). Purity was confirmed by staining with mAb to CD11c, CD14, HLA-DR, CD123, BDCA-2, and BDCA-4 (all from BD Pharmingen) and was typically >90%. Human NK cells were isolated by depletion from PBMC using NK cell isolation kit II according to the manufacturer’s information (Miltenyi Biotec). Purity was analyzed by staining with mAbs to CD56, CD3, CD14, and CD19 (all from BD Pharmingen). Human mDCs were isolated from PBMCs by CD1c (BDCA-1) and CD141 (BDCA-3) immunomagnetic bead positive selection according to manufacturer’s protocol (Miltenyi Biotec). Cell purity was controlled by BDCA-1, BDCA-3, CD14, and CD19 staining (all from BD Pharmingen). Isolated cells were plated on 96-well plates at 5 × 105 cells/ml and stimulated for 16 h with the indicated amount of ORNs complexed to DOTAP, R-848, CpG ODN, or DOTAP alone. After 16 h SN was harvested and cytokines or chemokines were measured.

For measurement of the expression of cell surface molecules, PBMCs or isolated cell subsets were incubated for 16 h. Cells were stained with mAbs for CD123, CD11c, and HLA-DR (pDCs) or CD14 (monocytes) in addition to mAbs for the indicated cell surface markers (BD Pharmingen). Expression of the different markers was measured by flow cytometry (FACSCalibur; BD Biosciences). Either the percentage of cells positive for a cell surface molecule or the mean expression was analyzed.

Sv129, C57BL/6 TLR7−/−, or C57BL/6 TLR9−/− mice (21) (6–8 wk of age) were used for the in vivo experiments, Sv129 mice were purchased from Charles River Canada and were housed in microisolators at the animal care facility of Coley Pharmaceutical. All studies were conducted under the approval of the institutional animal care committees and in accordance with the guidelines set forth by the Canadian Council on Animal Care. ORNs formulated with DOTAP were administered i.v. to the mice (n = 4 or 5 per group) and 3 h postinjection the animals were bled and IP-10, TNF-α, or IL-12 levels in plasma were measured by ELISA (BD Pharmingen).

Murine splenic DCs were derived from Sv129 mice using CD11c+ immunomagnetic bead positive selection according to manufacturer’s recommendations (Miltenyi Biotech). Murine cells were added to 96-well round-bottom plates with or without the addition of reagents as indicated. Culture SN were collected at 24 h and cytokine amounts were assessed using a commercially available ELISA kit (BD Pharmingen) or an in-house murine IFN-α ELISA developed using commercially available Abs (from PBL Biomedical Laboratories).

The Sprague-Dawley rats used for all in vivo experiments were purchased from Charles River Canada and housed in microisolators at the animal care facility of Coley Pharmaceutical. Splenocytes from Sprague-Dawley rats were pooled from three different rats, plated at 3 × 106 cells/ml, and incubated for 20 h with the indicated amounts of ORNs complexed to DOTAP, R-848 or DOTAP alone. SNs were harvested and used for commercially available ELISA (BD Pharmingen).

Blood was collected from the jugular vein by venipuncture using a 50-ml syringe containing 2 ml of 7.5% EDTA. The blood was centrifuged at 1400 × g for 20 min and the white blood cell-containing buffy coat was removed and resuspended in PBS (10 mM at pH 7.4) containing 0.1% EDTA. PBMC were obtained by overlaying the buffy coat on 54% Percoll (Pharmacia Biotech) and conducting centrifugation at 2,000 × g for 20 min. The PBMCs were then subjected to three washes using PBS (containing 0.1% EDTA). Stimulation of PBMCs was performed in 96-well, round-bottom plates (Nunc) using AIM-V medium supplemented with 2% FBS, 50 μg/ml streptomycin sulfate, 10 μg/ml gentamicin sulfate, 2 mM l-glutamine, and 50 μM 2-ME (Sigma-Aldrich). Cells (5 × 105) were cultured with ORN in triplicate wells in a 200-μl total volume and incubated for up to 48 h at 37°C in an atmosphere of 5% CO2 and 95% humidity. Culture SNs were collected and assayed for cytokine production.

Multiple sequence alignments were performed using ClustalW analysis (www.ebi.ac.uk/clustalw/) according to web site descriptions.

Previous studies suggested sequence-dependent recognition of short synthetic single-stranded U-rich RNAs by human TLR7 and TLR8 or mouse TLR7 (4, 19, 20, 21), although some reports argued against sequence-specific effects (7, 22). However, by comparing a phosphorothioate GU-rich (R-1075) sequence previously identified as a virus (HIV) genome-derived TLR7/8 ligand (4) to a homopolymeric phosphorothioate or phosphodiester poly(U) RNA sequence of the same length (18-mer), we found that the GU-rich ORN induced substantially stronger IFN-α production from human PBMCs (Fig. 1,A). This was evidence in favor of a sequence-specific difference in the TLR7-dependent immune response, and we further explored the impact of sequence composition on cellular activation. To enable us to differentiate between TLR7- and/or TLR8-mediated effects, we started to study the induction of IFN-α (from TLR7-positive cells) vs TNF-α (from TLR8-positive cells) cytokine release upon the stimulation of human PBMCs with ORNs (28, 29). To observe subtle influences of sequence modifications on cytokine responses, we first determined the minimal sequence motif contained in ORN R-1075 required to stimulate immune effects (Fig. 1,B). TNF-α and IFN-α production was analyzed upon stimulation with ORN R-1075, its central GU-rich sequence GUUGUGU, and different length variants of this sequence. Both the central sequence and its length variants were nested in either a random nonstimulatory DNA sequence or a nonstimulatory RNA sequence lacking U nucleotides (R-1263; see Table I). At least four nucleotides, UUGU, found in the GU-rich region (Fig. 1, B and C) were required to stimulate cytokine responses, whereas shorter RNA sequences like UUG, UG, or GU induced decreased cytokine stimulation. A combination of deoxyribonucleotides and ribonucleotides in the same molecule appeared to interfere with the RNA-mediated immune stimulatory effects, because nesting GU-rich RNA sequences in the random DNA sequence resulted in weaker cytokine release as compared with using the nonstimulatory RNA sequence surrounding the GU sequences (compare B and C in Fig. 1).

FIGURE 1.

The strongest IFN-α release needs a minimum of four defined ribonucleotides. Human PBMC were stimulated with 2.0 μM ORN complexed to 25 μg/ml DOTAP or with 25 μg/ml DOTAP alone (A). After 16 h SN was harvested and IFN-α and TNF-α were measured. Sequence variants of GUUGUGU were nested within a 20-mer random nonstimulatory ODN sequence (B) or nonstimulatory ORN sequence (R-1263) (C). Sequences are listed in Table I. Data shown are mean ± SEM of three donors. One of three independent experiments is shown. PO, Phosphodiester; PS, phosphorothioate.

FIGURE 1.

The strongest IFN-α release needs a minimum of four defined ribonucleotides. Human PBMC were stimulated with 2.0 μM ORN complexed to 25 μg/ml DOTAP or with 25 μg/ml DOTAP alone (A). After 16 h SN was harvested and IFN-α and TNF-α were measured. Sequence variants of GUUGUGU were nested within a 20-mer random nonstimulatory ODN sequence (B) or nonstimulatory ORN sequence (R-1263) (C). Sequences are listed in Table I. Data shown are mean ± SEM of three donors. One of three independent experiments is shown. PO, Phosphodiester; PS, phosphorothioate.

Close modal

Our data further suggest differences in the strength of IFN-α and TNF-α induction (Fig. 1 C). Whereas the 4-mer and 3-mer sequences UUGU and UUG showed only a moderate decrease of IFN-α compared with TNF-α secretion, the 2-mer sequences appeared to selectively induce TNF-α over IFN-α production. Therefore, at least one (for TNF-α production) or two U nucleotides (for TNF-α and IFN-α production) are needed to stimulate cytokine induction, and TLR7 stimulation (IFN-α production) appears more sensitive than TLR8 stimulation (TNF-α production) to subtle sequence changes.

To investigate the impact of sequence modifications on the stimulation of IFN-α and TNF-α production in more detail, 40 phosphorothioate ORNs differing in the U-containing 4-mer sequence derived from R-1075 were tested (Fig. 2 and data not shown). Surprisingly, specific sequence alterations converted an ORN from an IFN-α and TNF-α inducer to a relatively exclusive TNF-α inducer (Fig. 2). ORNs containing A in combination with U (referred here to as AU-rich or TLR8 ORNs) such as UAUU, UUAU, AUGU, AUAU, UAUA, UAUC, AUAC, CUAC, GUAC, CUAA, or AUAA still produced high amounts of TNF-α but little to no IFN-α (Fig. 2), while ORNs containing at least two U combined with G (or C) nucleotides (referred here to as GU-rich or TLR7/8 ORNs) like UUUU, GUUC, UUGU, GUUU, UUUC, UGUU, or UCUC induced high TNF-α and IFN-α.

FIGURE 2.

Differential IFN-α and TNF-α stimulation upon sequence modification. Human PBMCs were stimulated with 0.6 μM ORN complexed to DOTAP (10 μg/ml). ORN contain 4-mer sequence alterations of control ORN 1 or 2 (Table I). After 16 h, SN was harvested and IFN-α and TNF-α were measured. Data shown are mean (±SEM) of three donors.

FIGURE 2.

Differential IFN-α and TNF-α stimulation upon sequence modification. Human PBMCs were stimulated with 0.6 μM ORN complexed to DOTAP (10 μg/ml). ORN contain 4-mer sequence alterations of control ORN 1 or 2 (Table I). After 16 h, SN was harvested and IFN-α and TNF-α were measured. Data shown are mean (±SEM) of three donors.

Close modal

We further investigated whether the activity of ORNs containing such divergent AU or GU sequences may be altered and/or increased by nesting the identified 4-mer sequences UAUU and UUGU (strongest as 4-mers to induce TNF-α or TNF-α and IFN-α, respectively) in a longer 20-mer ORN (Table I, Fig. 3, and data not shown). All GU-rich ORNs (R-0005 to R-0008; Fig. 3,B) stimulated high IFN-α and TNF-α production, although differences in the potency to induce IFN-α secretion were observed. For example, whereas ORN R-0006 appeared to induce slightly more potent IFN-α induction than the other GU-rich ORNs, ORN R-0005 lacking the UUGU motif was a less potent inducer of IFN-α secretion. In contrast, an ORN containing a homopolymeric poly(U) sequence stimulated substantially weaker IFN-α cytokine induction, and a poly(G) as well as a poly(A) or poly(C) RNA sequence did not induce an effect (data not shown). The spacing also had substantial impact on AU-rich ORNs (Fig. 3,A). Whereas ORNs R-0003 and R-0004 with three or four U nucleotides bridging the A nucleotides were able to induce both TNF-α and IFN-α (resulting in principle in AUUU, UUUA, or UUUU motifs that stimulate both cytokines; Fig. 2), ORNs R-0001 and R-0002 with only one or two bridging U nucleotides stimulated TNF-α production but lacked IFN-α stimulation, indicating that the strength of the response depends on the U spacing between the motifs. Similar results as those for R-0006 and R-0002 were obtained with the same sequences in a phosphodiester backbone (data not shown).

FIGURE 3.

GU- and AU-rich ORNs show substantial differences between IFN-α and TNF-α release. Human PBMC were stimulated with the indicated concentrations of AU-rich (A) or GU-rich (B) ORN complexed to DOTAP. After 16 h, SN was harvested and IFN-α and TNF-α were measured. Data shown are mean ± SEM of three donors.

FIGURE 3.

GU- and AU-rich ORNs show substantial differences between IFN-α and TNF-α release. Human PBMC were stimulated with the indicated concentrations of AU-rich (A) or GU-rich (B) ORN complexed to DOTAP. After 16 h, SN was harvested and IFN-α and TNF-α were measured. Data shown are mean ± SEM of three donors.

Close modal

Further characterization of AU- and GU-rich ORN-mediated effects were performed using either HEK293 cells stably transfected with human TLR3, TLR7, TLR8, or TLR9 and an NF-κB luciferase reporter plasmid (Fig. 4, A–D). For these experiments, two representative GU- and AU-rich ORNs were used: R-0006 and R-0002 (see Fig. 3). Signaling in cells expressing TLR8 was induced by AU- and GU-rich ORNs in a sequence- and dose-dependent manner (Fig. 4, A and C), whereas only the GU-rich ORN stimulated signaling in TLR7-expressing cells (Fig. 4,D). ORNs not complexed to DOTAP (data not shown) or DOTAP alone did not induce an effect. Activation of immune cells by TLR ligands such as CpG ODN or ORN has been shown to require endosomal maturation and acidification (4, 19, 30). Inhibition of TLR8-mediated signaling induced by GU- and AU-rich ORNs as well as by R-848 with bafilomycin A1 resulted in loss of NF-κB activation (Fig. 4,B). Chloroquine strongly suppressed the ORN-mediated effects but did not affect the signaling induced by R-848, at least at the concentration used (Fig. 4 B). In data not shown, 10- to 100-fold higher concentrations of chloroquine were required to inhibit R-848.

FIGURE 4.

AU-rich and GU-rich ORNs stimulate human TLR7 and TLR8 signaling in a dose-dependent manner. A, HEK293 cells stably transfected with human TLR3, TLR8, or TLR9 containing a NF-κB-luciferase (NFkB-luc) reporter construct were incubated with the indicated ORN (10 μM complexed to 50 μg/ml DOTAP) or control stimuli (10 μM R-848, 50 μg/ml poly(rI:rC), 3.3 μM CpG ODN 10103, or 50 μg/ml DOTAP) for 16 h. NF-κB activation was measured by assaying luciferase activity. Results are given as fold induction above background (medium). B, Human TLR8-expressing HEK293 cells were incubated with the indicated ORN (15 μM complexed to 75 μg/ml DOTAP) or control stimuli (15 μM R-848 or 75 μg/ml DOTAP) together with 200 nM bafilomycin (Baf.) or 1 mM chloroquine (CQ) for 16 h. C, Human TLR8-expressing HEK293 cells were stimulated as in A. D, Human TLR7-expressing HEK293 cells containing a NF-κB-luciferase reporter construct were stimulated as in A. E, The mouse macrophage cell line Raw264.7 was stimulated with the indicated concentrations of ORN complexed to DOTAP as above, R-848, or DOTAP alone. After 16 h, SNs were harvested and cytokine release was measured. F, C57B/6 TLR7−/− or TLR9−/− mice (n = 4/group) were injected i.v. with 100 μg of ORN formulated with DOTAP or 100 μg of CpG ODN 1826 and bled after 3 h. IP-10 production was measured within plasma. Values are shown as pg/ml. G, HEK293 cells transiently transfected with murine TLR8 were stimulated for 16 h with the indicated concentrations of ORN complexed to DOTAP (100 μg/ml; 1/3 dilution), R-848, or DOTAP alone (100 μg/ml) in the presence of 5 μM poly(dT) (polyT) ODN. All data represent at least one of three or more independent experiments (for human PBMC, n = 3 donors).

FIGURE 4.

AU-rich and GU-rich ORNs stimulate human TLR7 and TLR8 signaling in a dose-dependent manner. A, HEK293 cells stably transfected with human TLR3, TLR8, or TLR9 containing a NF-κB-luciferase (NFkB-luc) reporter construct were incubated with the indicated ORN (10 μM complexed to 50 μg/ml DOTAP) or control stimuli (10 μM R-848, 50 μg/ml poly(rI:rC), 3.3 μM CpG ODN 10103, or 50 μg/ml DOTAP) for 16 h. NF-κB activation was measured by assaying luciferase activity. Results are given as fold induction above background (medium). B, Human TLR8-expressing HEK293 cells were incubated with the indicated ORN (15 μM complexed to 75 μg/ml DOTAP) or control stimuli (15 μM R-848 or 75 μg/ml DOTAP) together with 200 nM bafilomycin (Baf.) or 1 mM chloroquine (CQ) for 16 h. C, Human TLR8-expressing HEK293 cells were stimulated as in A. D, Human TLR7-expressing HEK293 cells containing a NF-κB-luciferase reporter construct were stimulated as in A. E, The mouse macrophage cell line Raw264.7 was stimulated with the indicated concentrations of ORN complexed to DOTAP as above, R-848, or DOTAP alone. After 16 h, SNs were harvested and cytokine release was measured. F, C57B/6 TLR7−/− or TLR9−/− mice (n = 4/group) were injected i.v. with 100 μg of ORN formulated with DOTAP or 100 μg of CpG ODN 1826 and bled after 3 h. IP-10 production was measured within plasma. Values are shown as pg/ml. G, HEK293 cells transiently transfected with murine TLR8 were stimulated for 16 h with the indicated concentrations of ORN complexed to DOTAP (100 μg/ml; 1/3 dilution), R-848, or DOTAP alone (100 μg/ml) in the presence of 5 μM poly(dT) (polyT) ODN. All data represent at least one of three or more independent experiments (for human PBMC, n = 3 donors).

Close modal

Additional proof for the selectivity of the AU- vs GU-rich ORN immune response comes from experiments on murine immune cells that express functional TLR7 but lack proper TLR8-dependent activation (8, 31). To investigate whether GU- and AU-rich ORNs differ in their capacity to induce murine TLR7 stimulation, the TLR7-expressing mouse macrophage cell lines RAW264.7 and J774 were stimulated with increasing concentrations of GU- and AU-rich ORN R-0006 and R-0002 (Fig. 4,E and data not shown). Only the GU-rich ORN stimulated TNF-α production in both macrophage cell lines, whereas the AU-rich ORN completely lacked an immune stimulatory effect. TLR7 specificity of the GU-rich ORN was also demonstrated by using splenocytes from TLR7−/− mice that did not respond to GU-rich ORN stimulation (Fig. 4 F).

Recently, activation of murine TLR8 was demonstrated by a combination of imidazoquinoline small molecules and homopolymer poly(dT) ODN (31). A poly(dT) ODN combined with either ORN R-0002 or R-0006 on HEK-293T cells transiently transfected with murine TLR8 resulted in substantial TLR8-mediated NF-κB signaling by both ORNs (Fig. 4 G). No response was observed for ORNs and R-848 in the absence of poly(dT) ODN (data not shown). Therefore, although the AU-rich ORN did not stimulate murine TLR7, TLR8 signaling was achieved for both ORN classes when combined with a homopolymer poly(dT) ODN similar to the results observed for human TLR8.

Human TLR7 and TLR9 colocalize in the same subpopulations of immune cells, B cells, and pDCs, whereas human TLR8 is strongly expressed in the myeloid compartment, e.g., monocytes, mDCs, or monocyte-derived DCs (15, 16, 17, 32, 33). Although our data strongly suggest that AU-rich ORNs do not signal through TLR7, only one major difference was observed between AU-rich and GU-rich ORNs when evaluating cytokine production in human PBMC cultures: the TLR8 ORN R-0002 induced minimal IFN-α or IFN-α related effects compared with the TLR7/8 ORN R-0006 (Table II). These data suggest that the main effects induced upon RNA TLR7 triggering in human immune cells are IFN-α and probably IFN-dependent effects.

Table II.

TLR8 ORNs lack IFN-α production from PBMC and pDCsa

PBMCPDCMonocytesmDC
R-0006R-0002R-0006R-0002R-0006R-0002R-0006R-0002
MeanSDMeanSDMeanSDMeanSDMeanSDMeanSDMeanSDMeanSD
IL-1Ra 79,575 5,915 57,522 7,282 66 56 −1 
MIP-1β 54,803 11,617 41,704 5,410 3,621 1,892 103 86.478 5,755 1,861 4,759 1,397 968 606 −65 
IL-6 54,510 11 51,505 1,293 350 232 12 15.565 10,646 2,485 11,871 728 1,660 211 1,149 168 
TNF-α 52,524 17,313 50,402 5,202 563 87 25 25.512 19,263 1,917 18,786 12,217 1,842 328 849 205 
IL-1 32,217 2,217 38,606 1,903 −1 −1 2,431 252 2,699 295 234 24 184 13 
MIP-1α 31,452 11,696 10,488 951 1,779 865 20 1.449 12,611 3,346 9,693 914 1,641 −47 1,315 −12 
IL-12p40 16,792 8,648 6,060 2,605 37 18 16 0.6997 348 51 225 15 256 133 94 46 
MCP-1 12,496 1,364 20,633 2,809 111 42 38 19.266 23 199 178 20 13 
IL-8 12,075 3,363 9,785 426 33,385 11,334 7,087 3,589.1 24,194 5,920 18,269 6,952 1,740 −283 1,581 −321 
IFN-α 5,065 364 10 1,295 300 −1.923 10 
RANTES 3,773 2,345 2,030 861 −1 −1 −1 −1 
IL-10 1,428 367 683 99 −2 −3 −2 −2.92 
IFN-γ 1,293 786 1,460 878 −1 −1 −1 
IP-10 1,278 136 1,066 376 −1.17 −1 −1 
IL-2R 745 179 842 37 168 102 −1 28 −2 24 37 
GM-CSF 675 39 688 74 −1 −1 158 187 31 
IL-15 523 62 512 23 −1 −1 −5 −2 
MIG-1 253 −23 103 −1 −1 
IL-7 235 36 204 −1 −1 
IL-13 93 89 −1 −1 22 26 
Eotaxin 18 21 −1 −1 
IL-4 13 12 −2 −2.04 
IL-5 13 13 −1 −1 
IL-2 −1 −1 
IL-17 −1 −1 
PBMCPDCMonocytesmDC
R-0006R-0002R-0006R-0002R-0006R-0002R-0006R-0002
MeanSDMeanSDMeanSDMeanSDMeanSDMeanSDMeanSDMeanSD
IL-1Ra 79,575 5,915 57,522 7,282 66 56 −1 
MIP-1β 54,803 11,617 41,704 5,410 3,621 1,892 103 86.478 5,755 1,861 4,759 1,397 968 606 −65 
IL-6 54,510 11 51,505 1,293 350 232 12 15.565 10,646 2,485 11,871 728 1,660 211 1,149 168 
TNF-α 52,524 17,313 50,402 5,202 563 87 25 25.512 19,263 1,917 18,786 12,217 1,842 328 849 205 
IL-1 32,217 2,217 38,606 1,903 −1 −1 2,431 252 2,699 295 234 24 184 13 
MIP-1α 31,452 11,696 10,488 951 1,779 865 20 1.449 12,611 3,346 9,693 914 1,641 −47 1,315 −12 
IL-12p40 16,792 8,648 6,060 2,605 37 18 16 0.6997 348 51 225 15 256 133 94 46 
MCP-1 12,496 1,364 20,633 2,809 111 42 38 19.266 23 199 178 20 13 
IL-8 12,075 3,363 9,785 426 33,385 11,334 7,087 3,589.1 24,194 5,920 18,269 6,952 1,740 −283 1,581 −321 
IFN-α 5,065 364 10 1,295 300 −1.923 10 
RANTES 3,773 2,345 2,030 861 −1 −1 −1 −1 
IL-10 1,428 367 683 99 −2 −3 −2 −2.92 
IFN-γ 1,293 786 1,460 878 −1 −1 −1 
IP-10 1,278 136 1,066 376 −1.17 −1 −1 
IL-2R 745 179 842 37 168 102 −1 28 −2 24 37 
GM-CSF 675 39 688 74 −1 −1 158 187 31 
IL-15 523 62 512 23 −1 −1 −5 −2 
MIG-1 253 −23 103 −1 −1 
IL-7 235 36 204 −1 −1 
IL-13 93 89 −1 −1 22 26 
Eotaxin 18 21 −1 −1 
IL-4 13 12 −2 −2.04 
IL-5 13 13 −1 −1 
IL-2 −1 −1 
IL-17 −1 −1 
a

Human PBMCs, purified pDCs, monocytes, or mDCs were stimulated with 1.0 μM (PBMC, pDCs, monocytes) or 2.0μM (mDCs) ORN complexed to DOTAP (10 μg/ml). After 16 h SNs were harvested and cytokines were measured by Luminex technology. Data shown are mean ± SEM of maximal activities (pg/ml) of two (PBMC and monocytes) or three (pDCs and mDCs) donors. DOTAP background values were subtracted from mean maximal activities and SEM.

To further analyze motif- and cell type-specific responses stimulated by TLR7/8 and TLR8 ORNs, pDCs, monocytes, mDCs, and NK cells were purified from human PBMCs. The AU-rich ORN R-0002 completely failed to induce IFN-α secretion from human pDCs, confirming our previous observation (Fig. 5). In contrast, monocytes, mDCs, and NK cells showed somewhat comparable TNF-α, IL-12p40, and IFN-γ cytokine release for both ORNs (Fig. 5). The TLR9 ligand C-Class CpG ODN 2395 exhibited a different cytokine pattern (34). We further investigated the release of a variety of cytokines and chemokines induced from purified pDCs, monocytes, and mDCs by the TLR8 and TLR7/8 ORNs (Table II). The data confirm the above observation that no IFN-α induction was observed in pDCs with the TLR8 ORN, whereas only pDCs responded to TLR7/8 ORN stimulation by IFN-α secretion. Stimulation of most other cytokines or chemokines from pDCs, monocytes, or mDCs appeared similar for both ORN classes.

FIGURE 5.

Cytokine secretion from purified human immune cells stimulated by TLR7/8 and TLR8 ORNs. CD123+ enriched cells (A), CD14+ enriched cells (B), mDCs (C), NK cells (D), or human PBMC (E–H) were stimulated with the indicated ORN at 2.0 μM complexed to DOTAP (20 μg/ml), 2.0 μM R-848, 0.5 μM ODN 2395, or DOTAP alone (20 μg/ml). After 16 h, SNs were harvested and cytokine production was measured. CD123+ purified pDC (I) or CD14+ purified monocytes (J) were incubated with the indicated amounts of ORN complexed to DOTAP or DOTAP alone. After 16 h, the cells were harvested and stained with CD123, CD11c, and HLA-DR (I) or CD14 and CD19 (J). Expression of CD86 or CD80 was acquired by flow cytometry. Data shown are mean ± SEM of three blood donors.

FIGURE 5.

Cytokine secretion from purified human immune cells stimulated by TLR7/8 and TLR8 ORNs. CD123+ enriched cells (A), CD14+ enriched cells (B), mDCs (C), NK cells (D), or human PBMC (E–H) were stimulated with the indicated ORN at 2.0 μM complexed to DOTAP (20 μg/ml), 2.0 μM R-848, 0.5 μM ODN 2395, or DOTAP alone (20 μg/ml). After 16 h, SNs were harvested and cytokine production was measured. CD123+ purified pDC (I) or CD14+ purified monocytes (J) were incubated with the indicated amounts of ORN complexed to DOTAP or DOTAP alone. After 16 h, the cells were harvested and stained with CD123, CD11c, and HLA-DR (I) or CD14 and CD19 (J). Expression of CD86 or CD80 was acquired by flow cytometry. Data shown are mean ± SEM of three blood donors.

Close modal

Because pDCs did not release IFN-α upon stimulation with the TLR8 ORN, we also investigated the effect of GU- and AU-rich ORNs on the up-regulation of cell surface activation markers on purified pDCs and monocytes (Fig. 5, I and J). The TLR8 ORN R-0002 did not induce CD86 expression upon culture with pDCs but up-regulated CD80 on purified monocytes. In contrast, the TLR7/8 ORN R-0006 stimulated CD86 and CD80 expression on pDCs and monocytes. Taken together, both classes of ORN ligands are capable of activating TLR8-expressing human monocytes, mDCs, and NK cells, whereas only GU-rich ORNs are capable of stimulating TLR7-expressing pDCs. In previous studies, human NK cells were demonstrated as expressing TLR8 and probably low levels of TLR7 and were the only cell type responsible for IFN-γ production upon stimulation with a TLR7/8 or TLR8 ligand, indicating direct stimulation of NK cells by ORN (21, 32, 35, 36), although one report suggested that the effects on NK cells were indirect (37).

Additional experiments aimed at evaluating the activity of the two different ORN classes on primary murine cells as well as in vivo. Results similar to the mouse macrophage cell lines were obtained in vitro; splenocytes and CD11c+ cells generated from Sv129 mice released a variety of cytokines and chemokines such as IL-12p40, IFN-α, IP-10, TNF-α, or IL-12p70 upon stimulation with the GU-rich ORN R-0006 or R-848, but not upon stimulation with the AU-rich ORN R-0002 (Fig. 6, A–C, and data not shown). In addition, in vivo studies upon i.v. administration of both ORNs complexed to DOTAP in Sv129 mice resulted in similar findings. Strong cytokine responses were detected for the TLR7/8 ligands R-848 and R-0006, but the TLR8 ligand R-0002 failed to stimulate a cytokine response (Fig. 6, D–I).

FIGURE 6.

TLR8 ORNs do not stimulate mouse immune responses. A–C, CD11c+ cells generated from Sv129 mice were stimulated with the indicated concentrations of ORN complexed to DOTAP (50 μg/ml), R-848, or DOTAP alone (50 μg/ml). After 20 h, SN were harvested and cytokines were measured. Data represent one from at least three independent experiments. D–I, Sv129 mice (n = 5/group) were injected i.v. with 3, 10, or 30 μg of ORN formulated with DOTAP or R-848 and bled after 3 h. Cytokine production was measured within blood serum. Values are shown as pg/ml.

FIGURE 6.

TLR8 ORNs do not stimulate mouse immune responses. A–C, CD11c+ cells generated from Sv129 mice were stimulated with the indicated concentrations of ORN complexed to DOTAP (50 μg/ml), R-848, or DOTAP alone (50 μg/ml). After 20 h, SN were harvested and cytokines were measured. Data represent one from at least three independent experiments. D–I, Sv129 mice (n = 5/group) were injected i.v. with 3, 10, or 30 μg of ORN formulated with DOTAP or R-848 and bled after 3 h. Cytokine production was measured within blood serum. Values are shown as pg/ml.

Close modal

Protein sequence comparison of TLR8 among different vertebrates (human, monkey, chimpanzee, dog, cattle, porcine, mouse, and rat) revealed few insertions or deletions across all LRRs. However, the strongest differences are detected in LRR3 (Table III). Although the human, chimpanzee, and monkey sequences are highly conserved, the rat, mouse, porcine, and bovine sequences exhibit deletions of four aa residues at position 106 (mouse), 103 (rat), or 102 (porcine) or an insertion of two aa at positions 105–106 (cattle). Interestingly, the porcine and bovine sequences have another deletion of two aa in the same region at position 97. Rat splenocytes are not responsive to the TLR8 ORN ligand R-0002 as observed for mouse immune cells (TNF-α, IL-6, IFN-γ, and IL-12; Fig. 7,A and data not shown). Similar data were observed for porcine PBMC; while the TLR7/8 ORN ligand R-0006 induced high levels of IL-12, IFN-α, TNF-α, and IFN-γ, ORN R-0002 failed to stimulate cytokines (Fig. 7,B and data not shown). In contrast, bovine PBMCs displayed responses to both ORNs (Fig. 7 C). One recent study suggested certain chemically modified nucleic acids to be TLR7/8 or TLR8 stimulators and demonstrated that both compounds activate immune cells in nonhuman primates in a manner similar to their effects on human immune cells (38). These data suggest that the LRR3 region may play an important role in the species-specific recognition of or signaling by single-stranded RNA.

Table III.

Multiple sequence alignment of TLR8 LRR3 of different speciesa

SpeciesAmino Acid SequenceAmino Acid Positions
Human PNVQHQNGNPGIQSNGLNITDGAFLNLKNLRELLLEDN 95–132 
Chimpanzee PNVQHQNGNPGIQSNGLNITDGAFLNLKNLRELLLEDN 116–153 
Monkey PNVQRQNGNPGMQSNGLNITDGAFLNLKNLRELLLEDN 95–132 
Dog ANPQHLSENPDN-KNGMNITDGAFLNLQNLNQLLLEDN 95–131 
Pig AKLW--PQS----ENGMTITDGAFLNLHHLRELLLEDN 95–126 
Cattle AKSQ--SGNPAV-KKAMTITDGAFLNLKHLRELLLEDN 95–129 
Rat AKQQHPNEN----KNGMNITEGAFLSLRNLTELLLEDN 95–128 
Mouse AKQQHPNEN----KNGMNITEGALLSLRNLTVLLLEDN 98–131 
SpeciesAmino Acid SequenceAmino Acid Positions
Human PNVQHQNGNPGIQSNGLNITDGAFLNLKNLRELLLEDN 95–132 
Chimpanzee PNVQHQNGNPGIQSNGLNITDGAFLNLKNLRELLLEDN 116–153 
Monkey PNVQRQNGNPGMQSNGLNITDGAFLNLKNLRELLLEDN 95–132 
Dog ANPQHLSENPDN-KNGMNITDGAFLNLQNLNQLLLEDN 95–131 
Pig AKLW--PQS----ENGMTITDGAFLNLHHLRELLLEDN 95–126 
Cattle AKSQ--SGNPAV-KKAMTITDGAFLNLKHLRELLLEDN 95–129 
Rat AKQQHPNEN----KNGMNITEGAFLSLRNLTELLLEDN 95–128 
Mouse AKQQHPNEN----KNGMNITEGALLSLRNLTVLLLEDN 98–131 
a

Multiple sequence alignment was performed using the ClustalW alignment program. The table shows the amino acids of LRR3.

FIGURE 7.

Purified rat splenocytes or porcine PBMC do not but bovine PBMC do respond to TLR8 ORNs by cytokine secretion. A, Splenocytes from Sprague-Dawley rats were pooled from three different rats and stimulated with the indicated ORN concentrations (complexed to 62.5 μg/ml DOTAP; 1/5 dilution), R-848, or DOTAP alone. SNs were harvested after 20 h and cytokines were measured. B and C, Porcine (B) and bovine PBMC (C) of five individual animals were stimulated with the indicated concentrations of ORN complexed to DOTAP (20 μg/m, 1/3 dilution). After 48 h, SNs were harvested and IFN-γ was measured.

FIGURE 7.

Purified rat splenocytes or porcine PBMC do not but bovine PBMC do respond to TLR8 ORNs by cytokine secretion. A, Splenocytes from Sprague-Dawley rats were pooled from three different rats and stimulated with the indicated ORN concentrations (complexed to 62.5 μg/ml DOTAP; 1/5 dilution), R-848, or DOTAP alone. SNs were harvested after 20 h and cytokines were measured. B and C, Porcine (B) and bovine PBMC (C) of five individual animals were stimulated with the indicated concentrations of ORN complexed to DOTAP (20 μg/m, 1/3 dilution). After 48 h, SNs were harvested and IFN-γ was measured.

Close modal

Phylogenetic analysis indicates that TLR7, 8, and 9 belong to one subfamily of TLRs and emerged as nearest neighbors. Within this subfamily, TLR7 and TLR8 are believed to build another small functional subfamily (1, 39). Previous data demonstrated that TLR7 is activated by genomic RNA from single-stranded RNA viruses (6, 7, 40) or synthetic RNA sequences (3, 4, 7, 19, 20, 21, 22, 41). Although some of these investigators did observe a lack of sequence specificity of TLR7-mediated responses, sequence-specific recognition of single-stranded RNA via human and murine TLR7 was suggested for sequences derived from single-stranded RNA viral genomes, U1 small nuclear RNA containing U and G nucleotides (4, 19), or certain small interfering RNA sequences (3, 41). Although all of these reports described sequence-specific effects of TLR7-dependent immune responses, no detailed analysis of the sequence selectivity of human TLR7 for single-stranded RNA was performed and, to our knowledge, no report investigated the potential sequence specificity of human TLR8 responses, although previously such differences were observed for small molecule TLR8 ligands (29).

Human TLR7 and TLR8 differ in their cell-type specific expression and, therefore, appear to activate divergent cell types resulting in differences in the cytokine patterns stimulated via each of these receptors. TLR7 agonists directly activate pDCs and induce IFN-α and IFN-α-regulated genes, whereas TLR8 agonists directly activate mDCs and monocytes and induce proinflammatory cytokines such as TNF-α and IL-12 (3, 4, 19, 21, 29). Target cell selectivity and diversity appear to be useful tools, in addition to recombinant cells expressing these receptors, to differentiate in vitro between potential TLR7, TLR7/8, and TLR8 RNA motifs. By comparing a variety of 4-mer RNA sequences on human primary immune cells for their potential to stimulate either IFN-α or TNF-α responses, we identified sequences stimulating both cytokines or only TNF-α, suggesting the existence of RNA motifs specific for both or the single receptors. The motif analysis defined specific GU-rich 4-mer sequences like UUGU, GUUC, GUUU, UUUC, UGUU, or UCUC activating human TLR7/8 by inducing IFN-α and proinflammatory cytokines and chemokines from cells expressing only TLR7 or both TLR7 and TLR8. In contrast, AU-rich sequences like AUGU, UAUA, AUAU, AUAC, UAUU, UUAU, CUAC, GUAC, or UAUC were found to induce the strongest TNF-α production by lacking substantial IFN-α secretion and revealed target cell and receptor selectivity by stimulating monocytes and mDCs, but not pDCs. Interestingly, our data also show that GU-rich ORNs induce substantially stronger IFN-α production from human PBMCs than a poly(U) ORN of the same length. Our motif study also suggests that human TLR7 is more sensitive to sequence modifications than human TLR8. Whereas nearly all tested RNA sequences still induced good to moderate TLR8 responses, human TLR7 appeared to prefer ORN sequences with at least two U ribonucleotides in combination with G and/or C. The selectivity for at least some similar RNA sequences may have evolutionary reasons; both TLRs lie in close vicinity to each other on the X chromosome, which suggests that they did arise from a tandem duplication of an ancestral gene. They contain two introns and three exons and more LRRs than most of the other TLRs (1, 24, 39), and the local gene order of TLR7 and TLR8 as well as the LRRs are preserved in humans and mice (42).

We investigated the sequence specificity of human TLR7 and TLR8 but did not determine in detail whether the murine TLR7 displays a comparable preference for GU-rich sequences. However, the murine TLR7 was stimulated only with a GU-rich but not an AU-rich ORN and the response required TLR7, as demonstrated by using splenocytes from TLR7-deficient mice. In contrast, AU-rich ORN in combination with a poly(dT) ODN did signal via murine TLR8. The failure of mouse TLR8 to signal upon encountering the “natural” human TLR8 ligand, single-stranded RNA, or small molecule TLR8 ligands alone (8, 31) may suggest that the activation of TLR8 requires additional factors or adaptors. In contrast to the combination of small molecule TLR8 ligands with poly(dT), AU-rich ORNs showed weaker murine TLR8-mediated NF-κB activation. This may be explained by ORNs and ODNs being both nucleic acids and, therefore, competing for the same uptake mechanism or TLR binding site in contrast to the small molecule ligands.

TLR7, 8, and 9 are expressed in intracellular endolysosomal compartments and chloroquine as well as bafilomycin A1 suppresses TLR7/8 and TLR8 ORN signaling via TLR8. Chloroquine previously was reported to interfere with the binding between TLR9 and CpG ODN (30), and because TLR9 is highly homologous to TLR7 and TLR8 it seems likely that chloroquine may also prevent a direct binding of the ORN to these other TLRs. Our results showing that chloroquine fails to efficiently block immune stimulation by small molecule ligands such as R-848 may point either to different TLR binding sites of R-848 vs nucleic acids and chloroquine or to differences in small molecule uptake and intracellular distribution.

Rat and mouse splenocytes as well as porcine PBMCs failed to respond to AU-rich ORNs, whereas GU-rich ORNs were capable to induce immune responses in cells of these species. In contrast, bovine PBMCs responded to both GU- and AU-rich ORNs. The cellular diversity of TLR7 and 8 within human, mouse, and rat appears not to be a good explanation for differences between GU-rich and AU-rich ORN; whereas TLR7 is expressed in human pDCs and B cells and TLR8 in monocytes and mDCs, TLR7 expression can be observed in mouse myeloid and lymphoid cells (8, 43, 44). More similar to human TLR7/8 expression, rat TLR7 can be found in pDCs and both rat TLR7 and 8 are found in monocytes (45, 46, 47). Although the multiple protein sequence alignment of TLR8 among different species indicates a high similarity, differences in terms of amino acid insertions or deletions are observed, especially in the LRR3. Deletions of four aa found in the mouse, rat, and porcine LRR3 sequences compared with the human sequence may explain the differences in TLR8 RNA sequence selectivity. The observed stimulation of bovine PBMC with both ORN classes may indicate the following: 1) that bovine TLR8 responds to TLR8-specific RNA ligands; 2) that the bovine TLR8 LRR3 sequence is different than all other species; or 3) that TLR7 substitutes for the activity of TLR8. Clearly, additional experiments are needed to further investigate the stimulation or lack of activity of TLR8 between different species. Structural analysis of TLR8 will probably result in more information of a potential ligand binding site and further help to explain functional differences among vertebrates. Initial mutagenesis studies identified regions in the human TLR8 extracellular domain most probably involved in signaling (48). Specific residues in LRR8 and LRR17 appear to be absolutely required for human TLR8 activation (48), and mutation of a residue in LRR17 of human TLR9 abolished specific binding to CpG DNA (30), indicating a role for LRR17 in ligand recognition for TLR7, 8, and 9 due to the similarities between these TLRs. Unfortunately, nothing is known yet about the potential function of LRR3, but our data suggest that this region plays a role in species-specific recognition of RNA by TLR8.

It is rather difficult to speculate about a potential evolutionary reason for GU vs AU sequence selectivity of TLR7 and TLR8. Highly conserved GU-containing RNA sequences can be found in the extracistronic 3′ terminal region of the genome of negative-strand viruses of the order Mononegavirales, including vesicular stomatitis virus, Sendai virus, and influenza virus (49). These regions are most critical for infection and the viral life cycle and contain highly conserved elements. Th1 cytokines are involved in the immune response to negative-strand viruses via TLR7 and TLR8 (5, 6, 50), and such cytokines can be found upon stimulation of human PBMCs with GU-rich ORNs derived from these sequences (21). In contrast, positive-strand alphaviruses contain repeated conserved sequence elements in the 3′ nontranslated region with AU-rich 4-mer motifs like AUGU, CUGA, AUAA, and AUGC (51). Alphaviruses are described as replicating rapidly in brain and spinal cord (52) and as being involved in experimental allergic encephalomyelitis (53). Th2 cytokines are produced in the CNS early upon experimental allergic encephalomyelitis infection before T cell infiltration, suggesting that the brain may be a site of preferential production of Th2 cytokines (52, 54), although IFN-α appears to play a role in the attenuation of alphaviruses (55). Innate immune receptors of the vertebrate immune system may have evolved to recognize specific pathogen RNA regions, resulting in differential patterns of immune responses giving rise to prevalent Th1 or Th2 responses. These responses may be at least in part driven by TLRs, leading to innate immune responses with strong (TLR7) or low (TLR8) type I IFN-dependent cytokine and chemokine production.

In summary, we have identified immune stimulatory ORN sequence motifs for human TLR7/8 and TLR8. AU-rich ORNs are TLR8 RNA ligands and mediate human TLR8 activation, whereas GU-rich ORNs are TLR7/8 RNA ligands and stimulate human TLR7 and TLR8 immune responses. However, besides the observed diverse immune effects and cellular specificities of the responses induced by the RNA TLR7/8 and TLR8 motifs, species-specific differences in the functionality of TLR8 were observed, and LRR3 in the ECD of TLR8 may play a role in this species specificity. The divergent immune modulatory responses of RNA TLR7/8 and TLR8 ligands suggests the potential of GU- and AU-rich ORNs to stimulate efficient TLR-dependent innate and adaptive immune effects that may be beneficial for different indications such as in cancer and allergic and infectious diseases.

We thank Dr. Bernhard Noll and Angela Lohner for performing quality control for identity and purity on the synthetic ODNs and ORNs. We are grateful to Dr. Shizuo Akira for providing us with TLR7- and TLR9-deficient mice.

All of the authors except G.K.M. are employees of Coley Pharmaceutical Group and may have a financial interest in the study’s outcome.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This study was supported by National Institute of Allergy and Infectious Diseases Contract HHSSN266200400044C.

3

Abbreviations used in this paper: ORN, oligoribonucleotide; BDCA, blood DC Ag; DC, dendritic cell; DOTAP, 1,3-dioleoyloxy-3-(trimethylammonium)propane; ECD, ectodomain; HEK, human embryonic kidney; IP-10, IFN-γ-inducible protein-10; LRR, leucine rich repeat; mDC, myeloid DC; ODN, oligodeoxynucleotide; PDC, plasmacytoid DC; poly(rI:rC), polyriboinosinic:polyribocytidylic acid; SN, supernatant.

1
Chuang, T. H., R. J. Ulevitch.
2000
. Cloning and characterization of a sub-family of human Toll-like receptors: hTLR7, hTLR8 and hTLR9.
Eur. Cytokine Network
11
:
372
-378.
2
Heil, F., P. Ahmad-Nejad, H. Hemmi, H. Hochrein, F. Ampenberger, T. Gellert, H. Dietrich, G. Lipford, K. Takeda, S. Akira, et al
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.
3
Hornung, V., M. Guenthner-Biller, C. Bourquin, A. Ablasser, M. Schlee, S. Uematsu, A. Noronha, M. Manoharan, S. Akira, A. de Fougerolles, et al
2005
. Sequence-specific potent induction of IFN-α by short interfering RNA in plasmacytoid dendritic cells through TLR7.
Nat. Med.
11
:
263
-270.
4
Heil, F., H. Hemmi, H. Hochrein, F. Ampenberger, C. Kirschning, S. Akira, G. Lipford, H. Wagner, S. Bauer.
2004
. Species-specific recognition of single-stranded RNA via toll-like receptor 7 and 8.
Science
303
:
1526
-1529.
5
Melchjorsen, J., S. B. Jensen, L. Malmgaard, S. B. Rasmussen, F. Weber, A. G. Bowie, S. Matikainen, S. R. Paludan.
2005
. Activation of innate defense against a paramyxovirus is mediated by RIG-I and TLR7 and TLR8 in a cell-type-specific manner.
J. Virol.
79
:
12944
-12951.
6
Lund, J. M., L. Alexopoulou, A. Sato, M. Karow, N. C. Adams, N. W. Gale, A. Iwasaki, R. A. Flavell.
2004
. Recognition of single-stranded RNA viruses by Toll-like receptor 7.
Proc. Natl. Acad. Sci. USA
101
:
5598
-5603.
7
Diebold, S. S., T. Kaisho, H. Hemmi, S. Akira, C. Reis e Sousa.
2004
. Innate antiviral responses by means of TLR7-mediated recognition of single-stranded RNA.
Science
303
:
1529
-1531.
8
Jurk, M., F. Heil, J. Vollmer, C. Schetter, A. M. Krieg, H. Wagner, G. Lipford, S. Bauer.
2002
. Human TLR7 or TLR8 independently confer responsiveness to the antiviral compound R-848.
Nat. Immunol.
3
:
499
9
Krieg, A. M., A. K. Yi, S. Matson, T. J. Waldschmidt, G. A. Bishop, R. Teasdale, G. A. Koretzky, D. M. Klinman.
1995
. CpG motifs in bacterial DNA trigger direct B-cell activation.
Nature
374
:
546
-549.
10
Hemmi, H., O. Takeuchi, T. Kawai, T. Kaisho, S. Sato, H. Sanjo, M. Matsumoto, K. Hoshino, H. Wagner, K. Takeda, S. Akira.
2000
. A Toll-like receptor recognizes bacterial DNA.
Nature
408
:
740
-745.
11
Bauer, S., C. J. Kirschning, H. Hacker, V. Redecke, S. Hausmann, S. Akira, H. Wagner, G. B. Lipford.
2001
. Human TLR9 confers responsiveness to bacterial DNA via species-specific CpG motif recognition.
Proc. Natl. Acad. Sci. USA
98
:
9237
-9242.
12
Iwasaki, A., R. Medzhitov.
2004
. Toll-like receptor control of the adaptive immune responses.
Nat. Immunol.
5
:
987
-995.
13
Ito, T., R. Amakawa, T. Kaisho, H. Hemmi, K. Tajima, K. Uehira, Y. Ozaki, H. Tomizawa, S. Akira, S. Fukuhara.
2002
. Interferon-α and interleukin-12 are induced differentially by Toll-like receptor 7 ligands in human blood dendritic cell subsets.
J. Exp. Med.
195
:
1507
-1512.
14
Krug, A., A. Towarowski, S. Britsch, S. Rothenfusser, V. Hornung, R. Bals, T. Giese, H. Engelmann, S. Endres, A. M. Krieg, G. Hartmann.
2001
. Toll-like receptor expression reveals CpG DNA as a unique microbial stimulus for plasmacytoid dendritic cells which synergizes with CD40 ligand to induce high amounts of IL-12.
Eur. J. Immunol.
31
:
3026
-3037.
15
Renn, C. N., D. J. Sanchez, M. T. Ochoa, A. J. Legaspi, C. K. Oh, P. T. Liu, S. R. Krutzik, P. A. Sieling, G. Cheng, R. L. Modlin.
2006
. TLR activation of Langerhans cell-like dendritic cells triggers an antiviral immune response.
J. Immunol.
177
:
298
-305.
16
Miettinen, M., T. Sareneva, I. Julkunen, S. Matikainen.
2001
. IFNs activate toll-like receptor gene expression in viral infections.
Genes Immun.
2
:
349
-355.
17
Zarember, K. A., P. J. Godowski.
2002
. Tissue expression of human Toll-like receptors and differential regulation of Toll-like receptor mRNAs in leukocytes in response to microbes, their products, and cytokines.
J. Immunol.
168
:
554
-561.
18
Bernasconi, N. L., N. Onai, A. Lanzavecchia.
2003
. A role for Toll-like receptors in acquired immunity: upregulation of TLR9 by BCR triggering in naive B cells and constitutive expression in memory B cells.
Blood
101
:
4500
-4504.
19
Vollmer, J., S. Tluk, C. Schmitz, S. Hamm, M. Jurk, A. Forsbach, S. Akira, K. M. Kelly, W. H. Reeves, S. Bauer, A. M. Krieg.
2005
. Immune stimulation mediated by autoantigen binding sites within small nuclear RNAs involves Toll-like receptors 7 and 8.
J. Exp. Med.
202
:
1575
-1585.
20
Sioud, M..
2006
. Single-stranded small interfering RNA are more immunostimulatory than their double-stranded counterparts: a central role for 2′-hydroxyl uridines in immune responses.
Eur. J. Immunol.
36
:
1222
-1230.
21
Forsbach, A., J. G. Nemorin, K. Volp, U. Samulowitz, C. Montino, C. Muller, S. Tluk, S. Hamm, S. Bauer, G. B. Lipford, J. Vollmer.
2007
. Characterization of conserved viral leader RNA sequences that stimulate innate immunity through TLRs.
Oligonucleotides
17
:
405
-418.
22
Diebold, S. S., C. Massacrier, S. Akira, C. Paturel, Y. Morel, C. R. Sousa.
2006
. Nucleic acid agonists for Toll-like receptor 7 are defined by the presence of uridine ribonucleotides.
Eur. J. Immunol.
36
:
3256
-3267.
23
Bell, J. K., G. E. Mullen, C. A. Leifer, A. Mazzoni, D. R. Davies, D. M. Segal.
2003
. Leucine-rich repeats and pathogen recognition in Toll-like receptors.
Trends Immunol.
24
:
528
-533.
24
Matsushima, N., T. Tanaka, P. Enkhbaya, T. Mikami, M. Taga, K. Ymada, Y. Kuroki.
2007
. Comparative sequence analysis of leucine-rich repeats (LRRs) within vertebrate toll-like receptors.
BMC Genomics
8
:
124
25
Jurk, M., A. Kritzler, B. Schulte, S. Tluk, C. Schetter, A. M. Krieg, J. Vollmer.
2006
. Modulating responsiveness of human TLR7 and 8 to small molecule ligands with T-rich phosphorothiate oligodeoxynucleotides.
Eur. J. Immunol.
36
:
1815
-1826.
26
Gorden, K. K., X. Qiu, J. J. Battiste, P. P. Wightman, J. P. Vasilakos, S. S. Alkan.
2006
. Oligodeoxynucleotides differentially modulate activation of TLR7 and TLR8 by imidazoquinolines.
J. Immunol.
177
:
8164
-8170.
27
Jurk, M., A. Kritzler, H. Debelak, J. Vollmer, A. M. Krieg, E. Uhlmann.
2006
. Structure-activity relationship studies on the immune stimulatory effects of base-modified CpG toll-like receptor 9 agonists.
ChemMedChem
1
:
1007
-1014.
28
Ghosh, T. K., D. J. Mickelson, J. C. Solberg, K. E. Lipson, J. R. Inglefield, S. S. Alkan.
2007
. TLR-TLR cross talk in human PBMC resulting in synergistic and antagonistic regulation of type-1 and 2 interferons, IL-12, and TNF-α.
Int. Immunopharmacol.
7
:
1111
-1121.
29
Gorden, K. B., K. S. Gorski, S. J. Gibson, R. M. Kedl, W. C. Kieper, X. Qiu, M. A. Tomai, S. S. Alkan, J. P. Vasilakos.
2005
. Synthetic TLR agonists reveal functional differences between human TLR7 and TLR8.
J. Immunol.
174
:
1259
-1268.
30
Rutz, M., J. Metzger, T. Gellert, P. Luppa, G. B. Lipford, H. Wagner, S. Bauer.
2004
. Toll-like receptor 9 binds single-stranded CpG-DNA in a sequence- and pH-dependent manner.
Eur. J. Immunol.
34
:
2541
-2550.
31
Gorden, K. K., X. X. Qiu, C. C. Binsfeld, J. P. Vasilakos, S. S. Alkan.
2006
. Cutting edge: activation of murine TLR8 by a combination of imidazoquinoline immune response modifiers and PolyT oligodeoxynucleotides.
J. Immunol.
177
:
6584
-6587.
32
Hornung, V., S. Rothenfusser, S. Britsch, A. Krug, B. Jahrsdorfer, T. Giese, S. Endres, G. Hartmann.
2002
. Quantitative expression of toll-like receptor 1–10 mRNA in cellular subsets of human peripheral blood mononuclear cells and sensitivity to CpG oligodeoxynucleotides.
J. Immunol.
168
:
4531
-4537.
33
Berkeredjian-Ding, I. B., M. Wagner, V. Hornung, T. Giese, M. Schnurr, S. Endres, G. Hartmann.
2005
. Plasmacytoid dendritic cells control TLR7 sensitivity of naive B cells via type I IFN.
J. Immunol.
174
:
4043
-4050.
34
Vollmer, J., R. Weeratna, P. Payette, M. Jurk, C. Schetter, M. Laucht, T. Wader, S. Tluk, M. Liu, H. L. Davis, A. M. Krieg.
2004
. Characterization of three CpG oligodeoxynucleotide classes with distinct immunostimulatory activities.
Eur. J. Immunol.
34
:
251
-262.
35
Hart, O. M., V. Athie-Morales, G. M. O’Connor, C. M. Gardiner.
2005
. TLR7/8-mediated activation of human NK cells results in accessory cell-dependent IFN-γ production.
J. Immunol.
175
:
1636
-1642.
36
Lauzon, N. M., F. Mian, R. Mackenzie, A. A. Ashkar.
2006
. The direct effects of Toll-like receptor ligands on human NK cell cytokine production and cytotoxicity.
Cell. Immunol.
241
:
102
-112.
37
Gorski, K. S., E. L. Waller, J. Bjornton-Severson, J. A. Hanten, C. L. Riter, W. C. Kieper, K. B. Gorden, J. S. Miller, J. P. Vasilakos, M. A. Tomai, S. S. Alkan.
2006
. Distinct indirect pathways govern human NK-cell activation by TLR-7 and TLR-8 agonists.
Int. Immunol.
18
:
1115
-1126.
38
Lan, T., E. R. Kandimalla, D. Yu, L. Bhagat, Y. Li, D. Wang, F. Zhu, J. X. Tang, M. R. Putta, Y. Cong, et al
2007
. Stabilized immune modulatory RNA compounds as agonists of Toll-like receptors 7 and 8.
Proc. Natl. Acad. Sci. USA
104
:
13750
-13755.
39
Du, X., A. Poltorak, Y. Wei, B. Beutler.
2000
. Three novel mammalian toll-like receptors: gene structure, expression, and evolution.
Eur. Cytokine Network
11
:
362
-371.
40
Barchet, W., A. Krug, M. Cella, C. Newby, J. A. Fischer, A. Dzionek, A. Pekosz, M. Colonna.
2004
. Dendritic cells respond to influenza virus through TLR7- and PKR-independent pathways.
Eur. J. Immunol.
35
:
236
-242.
41
Sioud, M..
2005
. Induction of inflammatory cytokines and interferon responses by double-stranded and single-stranded siRNAs is sequence-dependent and requires endosomal localization.
J. Mol. Biol.
348
:
1079
-1090.
42
Roach, J. C., G. Glusman, L. Rowen, A. Kaur, M. K. Purcell, K. D. Smith, L. E. Hood, A. Aderem.
2005
. The evolution of vertebrate Toll-like receptors.
Proc. Natl. Acad. Sci. USA
102
:
9577
-9582.
43
Edwards, A. D., S. S. Diebold, E. M. Slack, H. Tomizawa, H. Hemmi, T. Kaisho, S. Akira, C. Reis e Sousa.
2003
. Toll-like receptor expression in murine DC subsets: lack of TLR7 expression by CD8 α+ DC correlates with unresponsiveness to imidazoquinolines.
Eur. J. Immunol.
33
:
827
-833.
44
Applequist, S. E., R. P. Wallin, H. G. Ljunggren.
2002
. Variable expression of Toll-like receptor in murine innate and adaptive immune cell lines.
Int. Immunol.
14
:
1065
-1074.
45
Hubert, F. X., C. Voisine, C. Louvet, M. Heslan, R. Josien.
2004
. Rat plasmacytoid dendritic cells are an abundant subset of MHC class II+ CD4+CD11b OX62 and type I IFN-producing cells that exhibit selective expression of toll-like receptors 7 and 9 and strong responsiveness to CpG.
J. Immunol.
172
:
7485
-7494.
46
Hubert, F. X., C. Voisine, C. Louvet, J. M. Heslan, A. Ouabed, M. Heslan, R. Josien.
2006
. Differential pattern recognition receptor expression but stereotyped responsiveness in rat spleen dendritic cell subsets.
J. Immunol.
177
:
1007
-1016.
47
Hammerbeck, D. M., G. R. Burleson, C. J. Schuller, J. P. Vasilakos, M. Tomai, E. Egging, F. R. Cochran, S. Woulfe, R. L. Miller.
2007
. Administration of a dual Toll-like receptor 7 and Toll-like receptor 8 agonist protects against influenza in rats.
Antiviral Res.
73
:
1
-11.
48
Gibbard, R. J., P. J. Morley, N. J. Gay.
2006
. Conserved features in the extracellular domain of human toll-like receptor 8 are essential for pH-dependent signaling.
J. Biol. Chem.
281
:
27503
-27511.
49
Conzelmann, K. K..
1998
. Nonsegmented negative-strand RNA viruses: genetics and manipulation of viral genomes.
Annu. Rev. Genet.
32
:
123
-162.
50
Diebold, S. S., M. Montoya, H. Unger, L. Alexopoulou, P. Roy, L. E. Haswell, A. Al Shamkhani, R. Flavell, P. Borrow, C. Reis e Sousa.
2003
. Viral infection switches non-plasmacytoid dendritic cells into high interferon producers.
Nature
424
:
324
-328.
51
Strauss, J. H., E. G. Strauss.
1994
. The alphaviruses: gene expression, replication, and evolution.
Microbiol. Rev.
58
:
491
-562.
52
Rowell, J. F., D. E. Griffin.
1999
. The inflammatory response to nonfatal Sindbis virus infection of the nervous system is more severe in SJL than in BALB/c mice and is associated with low levels of IL-4 mRNA and high levels of IL-10-producing CD4+ T cells.
J. Immunol.
162
:
1624
-1632.
53
Eralinna, J. P., M. Soilu-Hanninen, M. Roytta, V. Hukkanen, A. A. Salmi, R. Salonen.
1996
. Blood-brain barrier breakdown and increased intercellular adhesion molecule (ICAM-1/CD54) expression after Semliki Forest (A7) virus infection facilitates the development of experimental allergic encephalomyelitis.
J. Neuroimmunol.
66
:
103
-114.
54
Wesselingh, S. L., B. Levine, R. J. Fox, S. Choi, D. E. Griffin.
1994
. Intracerebral cytokine mRNA expression during fatal and nonfatal alphavirus encephalitis suggests a predominant type 2 T cell response.
J. Immunol.
152
:
1289
-1297.
55
White, L. J., J. G. Wang, N. L. Davis, R. E. Johnston.
2001
. Role of αβ interferon in Venezuelan equine encephalitis virus pathogenesis: effect of an attenuating mutation in the 5′ untranslated region.
J. Virol.
75
:
3706
-3718.