TLR9 ligands induce S100A8 in macrophages via a STAT3-dependent pathway which requires IL-10 and PGE2

Abstract

S100A8 and S100A9 are highly-expressed calcium-binding proteins in neutrophils and monocytes, and in subsets of macrophages in inflammatory lesions. Unmethylated CpG motifs found in bacterial and viral DNA are potent activators of innate immunity via Toll-like receptor 9 (TLR9). S100A8, but not S100A9, mRNA and protein was directly induced by CpG-DNA in murine and human macrophages. Induction in murine macrophages peaked at 16 h. CpG-DNA-induced S100A8 required de novo protein synthesis; IL-10 and Prostaglandin E2 (PGE2) synergistically enhanced expression and promoted earlier gene induction. Inhibitors of endogenous IL-10, PGE2, and the E prostanoid (EP) 4 receptor strongly suppressed S100A8 expression, particularly when combined. Thus, S100A8 induction by E. coli DNA required both IL-10 and PGE2/EP4 signaling. The MAPKs, PI3K and JAK pathways were essential, whereas ERK1/2 appeared to play a direct role. S100A8 induction by CpG-DNA was controlled at the transcriptional level. The promoter region responsible for activation, either directly, or indirectly via IL-10 and PGE2, was located within a -178 to -34-bp region and required STAT3 binding. Because of the robust links connecting IL-10 and PGE2 with an anti-inflammatory macrophage phenotype, the induction profile of S100A8 strongly indicates a role for this protein in resolution of inflammation.

Figure 1. Induction of S100A8 mRNA in macrophages by E. coli DNA, M. luteus DNA, and synthetic CpG-ONDs.

(A) RAW 264.7 macrophages were incubated with E. coli DNA (3 µg/ml) or M. luteus DNA (3 µg/ml) ±0.3 ng/ml LPS. Controls include placental DNA (3 µg/ml) ± LPS (0.3 ng/ml), Turbo DNase-treated E. coli DNA or M. luteus DNA ± LPS (0.3 ng/ml). RAW cells were also stimulated with two doses of LPS alone as positive controls (0.3 or 100 ng/ml). (B) RAW cells were untreated (medium) or treated with a synthetic non-CpG-containing OND 1982 (nCpG), a murine-oriented CpG-containing ODN (CpG-1826), or a human-oriented CpG-containing ODN (CpG-2006) (all 3 µM final concentration) ± LPS (0.3 ng/ml). (C) medium control, CpG-1826 (3 µM) or LPS (20 ng/ml) were co-stimulated with inhibitory CpG ODN 2088 (iCpG) 2 or 6 µM. (D) Elicited murine macrophages were untreated or stimulated with human placental DNA or E. coli DNA at the indicated concentrations. (E) Human monocyte-derived macrophages stimulated with 3 µM of nCpG or CpG-2216; LPS (20 ng/ml) was positive control. Combined data from 3 independent experiments are presented as the mean ± SD (*, p<0.05). All treatments were for 24 h before cell harvest. Total RNAs were extracted and S100A8 mRNA quantitated by real-time RT-PCR and normalized to the house-keeping genes (murine HPRT was used as endogenous control for A, B and C; murine β-actin for D; human β-actin for E). All data, except E, are represented as means (relative to the HPRT or actin mRNA levels) ± SD of duplicate measurements and are representative of at least 3 experiments.

Figure 2. Induction of S100A8 by CpG-DNA is dose-dependent.

(A) RAW 264.7 macrophages were incubated with E. coli DNA at the indicated concentrations ±0.3 ng/ml LPS for 24 h before harvesting and (A) mRNAs quantitated by qRT-PCR with different doses of DNase-treated E. coli DNA ± LPS (0.3 ng/ml) as control. (B) S100A8 levels in culture supernates from (A) were quantitated by ELISA. Other controls included various concentrations of placental DNA ± LPS (0.3 ng/ml) or DNase-treated E. coli DNA, but mRNA levels were same as control and proteins were below the detectable level (not shown). (C) S100A8 in supernates from RAW cells treated with synthetic nCpG-1982 or CpG-1826 at the indicated concentrations was quantitated. QRT-PCR data represent means (relative to the γ-actin mRNA levels) ± SD of duplicates and are representative of 3 experiments. ELISA data are means ± SD of duplicate samples from 2 separate experiments.

Figure 3. Induction of S100A8 in macrophages by E. coli DNA is late and requires newly synthesis proteins.

(A–C) RAW264.7 macrophages were treated with 3 µg/ml placental DNA, 3 µg/ml E. coli DNA, or 20 ng/ml LPS for the indicated times. S100A8 (A), IL-10 (B), or COX-2 (C) mRNAs quantitated by qRT-PCR. (D) Effects of CHX (2 µg/ml) on LPS (20 ng/ml), E. coli DNA (3 µg/ml), or CpG-containing ODN (1826, 3 µM)-induced S100A8 mRNA expression. (E) Cells incubated with E. coli DNA (3 µg/ml) ± IL-10 (2 ng/ml) or ± PGE₂ (10 µM), harvested at the indicated time points and mRNAs quantitated. Data represent means ± SD of duplicates and representative of 3 experiments.

Figure 4. Induction of S100A8 in macrophages by CpG-ODN requires both IL-10 and PGE₂/EP4 dual signaling pathways.

(A and B) RAW cells were incubated with CpG-ODN (3 µM) for 20 h in the presence/absence of two doses (Lo: low, 1 ng/ml; Hi: high, 10 ng/ml) of IL-10 Ab and/or COX inhibitors (NS389: Lo, 5 µM, Hi, 50 µM; Indomethacin (Indo): Lo, 2 µM; Hi, 20 µM). Cells and supernates were harvested, and S100A8 mRNA (A) or protein (B) quantitated. Data represent means ± SD of duplicates and representative of 3 experiments. (C and D) RAW cells were incubated with CpG-ODN (3 µM) in the presence/absence of two doses (Lo, 1 µM; Hi, 10 µM) of antagonists for EP1 (SC-19220), EP2 (AH-6809), and EP4 (AH-23848) for 8 (C) and 20 h (D). IL-10 (C) and S100A8 mRNA (D) quantitated by qRT-PCR. (E and F) RAW cells were untreated or treated with CpG-ODN (3 µM) for 6 h in the presence/absence of anti-IL-10 Ab (5 ng/ml) and/or NS389 (25 µM), and IL-10 (E) and Cox2 (F) mRNA quantitated by qRT-PCR, For C–F, data represent means ± SD of −5 separate experiments. *P<0.05, ***P<0.001 of differences between samples treated with CpG-ODN alone and CpG-ODN plus various inhibitors.

Figure 5. Contribution of MAPKs, PI3K, and JAK2 to E. coli-DNA-mediated S100A8, IL-10 and COX-2 mRNA induction.

(A–C) RAW 264.7 cells were pretreated with DMSO (vehicle control), U0126 (2.5 µM), SB202190 (2.5 µM), JNK inhibitor (5 µM), LY209002 (10 µM), or AG490 (20 ng/ml) for 20 min then untreated or stimulated with E. coli DNA (3 µg/ml) for 20 h (A), or 4 h (B, C). mRNA for S100A8 (A), IL-10 (B) or COX-2 (C) was quantitated using qRT-PCR. (D) RAW cells were treated using same conditions as in (A), except that AG490 was replaced by Ruxolitinib (0.5 µM); IL-10 in supernatants were quantitated by ELISA. (E) RAW cells were treated with inhibitors for 20 min, stimulated with E. coli DNA for 30 min, and then ± PGE₂ (10 µM) or IL-10 (5 ng/ml) added and cells cultured for 20 h. S100A8 mRNA levels were quantitated; endogenous control was γ-actin. Data represent % S100A8 mRNA relative to E. coli DNA-stimulated samples without inhibitors ± SD of 3 separate experiments. (F) RAW 264.7 cells were transiently co-transfected with the S100A8 promoter (–178/465 bp) luciferase reporter construct and pRL-TK-luciferase control plasmid and pretreated with various MAPK and COX-2 inhibitors, or anti-IL-10 mAb for 20 min before adding E. coli DNA (3 µg/ml) for 16 h. Normalized S100A8 promoter-driven luciferase activities in cell extracts were analysed. Except for (E), data represent means ± SD of 3–4 separate experiments. *P<0.05, **P<0.01, ***P<0.001 of differences between E. coli DNA treatment alone and samples treated with E. coli DNA plus various inhibitors.

Figure 6. STAT3 is a critical transcription factor in CpG-DNA-induced S100A8 expression.

(A) RAW 264.7 cells were transiently co-transfected with pRL-TK-luciferase and full-length (−917/465) or a series of 5′-deleted S100A8 promoter-luciferase reporters for 24 h. pGL-basic and pGL-promoter plasmids were transfected as controls. Transfected cells were either left untreated (open bars) or stimulated with CpG-ODN (3 µM, solid bars) for 16 h. Luciferase activities in cell extracts were analysed; data represent means (% luciferase activity of pGL-promoter luciferase reporter in cells stimulated with CpG-ODN) ± SD of 3 separate experiments. (B) A potential STAT3-binding element overlaps the C/EBP binding site in both murine and human S100A8 promoter (shaded). The matched bases of STAT3 binding sites are compared with verified binding sites in the VEGF and Jab1 promoters. (C) RAW cells were treated with CpG-ODN (3 µM) for the indicated times (h) before harvest, then proteins in lysates separated by SDS-PAGE. Western blotting used a phosphor-specific anti-STAT3 antibody; stripped blots were re-probed with anti-STAT3 antibody. RAW cells treated with IL-10 for 30 min or IFNα-treated Hela (5 min) were positive controls. Results are representative of 3 experiments. (D) RAW 264.7 cells were transiently co-transfected with the S100A8 promoter (−178/465 bp) luciferase reporter construct and pRL-TK-luciferase control plasmid, with either empty vectors, dominant-positive (Stat3-C), or dominant-negative (Stat3-β) STAT3 constructs for 24 h. Cells were then stimulated with (□) or without (▪) CpG-ODN (3 µM) for 16 h prior to performing dual-luciferase assays. Results are mean ± SD of 3 independent experiments. *P<0.05 or **P<0.01 of differences with or without CpG-ODN treatments. (E) Time course of ChIP assays indicated binding of STAT3 to the S100A8 promoter in RAW cells stimulated with CpG-ODN (3 µM) (▪). The same region was not enriched by IgG control (□). (F) RAW cells were untreated or treated with 10 µM PGE₂ or 3 µM CpG-ODN for the indicated time and C/EBPβ mRNA quantitated by qRT-PCR. Results are means ± SD of 3 independent experiments. *P<0.05 or **P<0.01 of differences compared to medium control.