TLR2-dependent inhibition of macrophage responses to IFN-gamma is mediated by distinct, gene-specific mechanisms

Abstract

Mycobacterium tuberculosis uses multiple mechanisms to avoid elimination by the immune system. We have previously shown that M. tuberculosis can inhibit selected macrophage responses to IFN-gamma through TLR2-dependent and -independent mechanisms. To specifically address the role of TLR2 signaling in mediating this inhibition, we stimulated macrophages with the specific TLR2/1 ligand Pam(3)CSK(4) and assayed responses to IFN-gamma. Pam(3)CSK(4) stimulation prior to IFN-gamma inhibited transcription of the unrelated IFN-gamma-inducible genes, CIITA and CXCL11. Surface expression of MHC class II and secretion of CXCL11 were greatly reduced as well, indicating that the reduction in transcripts had downstream effects. Inhibition of both genes required new protein synthesis. Using chromatin immunoprecipitation, we found that TLR2 stimulation inhibited IFN-gamma-induced RNA polymerase II binding to the CIITA and CXCL11 promoters. Furthermore, TATA binding protein was unable to bind the TATA box of the CXCL11 promoter, suggesting that assembly of transcriptional machinery was disrupted. However, TLR2 stimulation affected chromatin modifications differently at each of the inhibited promoters. Histone H3 and H4 acetylation was reduced at the CIITA promoter but unaffected at the CXCL11 promoter. In addition, NF-kappaB signaling was required for inhibition of CXCL11 transcription, but not for inhibition of CIITA. Taken together, these results indicate that TLR2-dependent inhibition of IFN-gamma-induced gene expression is mediated by distinct, gene-specific mechanisms that disrupt binding of the transcriptional machinery to the promoters.

Figure 1. TLR2 stimulation inhibits IFN-γ-induced transcription of a subset of genes.

BALB/c BMDM were treated with 10 ng/ml Pam₃CSK₄ for 8 h, and then 20 ng/ml IFN-γ for 4, 8, and 12 h. Total RNA was harvested, reverse transcribed, and CXCL11 (A), CIITA (B), and NOS2 (C) expression analyzed by quantitative real-time PCR (qPCR). All values were normalized to GAPDH. Results are shown as fold induction compared to untreated sample without Pam₃CSK₄ or IFN-γ. *, p<0.01, **, p<0.05 comparing IFN-γ alone samples with those treated with Pam₃CSK₄ prior to IFN-γ (as determined by two-tailed t test). Results in (A) are expressed as means±SEM of three independent experiments. Results in (B) and (C) are representative of three independent experiments. Similar results were obtained with C57BL/6 BMDM.

Figure 2. TLR2-mediated inhibition of IFN-γ induction of CXCL11 and CIITA decreases expression of protein products.

A. BALB/c BMDM were treated with 10 ng/ml Pam₃CSK₄ for 8 h followed by 20 ng/ml IFN-γ for 4, 8, and 12 h. Culture supernatants were collected and assayed for CXCL11 protein by ELISA. *, p<0.01, **, p<0.05 comparing IFN-γ alone with Pam₃CSK₄ and IFN-γ treated samples (as determined by two-tailed t-test). B and C. BMDM were treated with 10 ng/ml Pam₃CSK₄ for 12–15 h prior to stimulation with 20 ng/ml IFN-γ for 24 h. Cells were stained with Alexa 647-conjugated anti-mouse I-A/I-E and analyzed by flow cytometry. Data shown are fluorescence intensity vs. cell number (B) and mean I-A/I-E fluorescence (C). Results are expressed as means±SEM from two independent experiments (A) and are representative of at least five independent experiments (B and C).

Figure 3. New protein synthesis is required for inhibition of CIITA and CXCL11.

BMDM were pretreated with 500 nM cycloheximide or DMSO for 1 h prior to treatment with 10 ng/ml Pam₃CSK₄ for 8 h followed by 20 ng/ml IFN-γ for 4 h in the continued presence or absence of inhibitor. Total RNA was harvested after IFN-γ stimulation. CXCL11 (A), CIITA (B), and NOS2 (C) expression was assayed by qPCR and normalized to GAPDH and untreated samples. The concentration of cycloheximide used inhibited TNF production (as a measure of protein synthesis) by over 90% with minimal cell death. Statistical significance between IFN-γ alone samples and those treated with Pam₃CSK₄ prior to IFNγ was determined by two-tailed t-test.

Figure 4. TLR2 stimulation prevents binding of general transcriptional machinery to the CIITA and CXCL11 promoters.

BMDM were treated with 10 ng/ml Pam₃CSK₄ for 8–9 h, then 20 ng/ml IFN-γ for 4 h. Cross-linked DNA was sheared and immunoprecipitated with anti-PolII (A) or anti-TBP (B) antibodies. Precipitated and input DNA for each sample were assayed by qPCR with primers specific for the transcriptional start site in the promoters of CXCL11, CIITA, and NOS2 (A) or the TATA box of the CXCL11 promoter (B). All values were normalized to GAPDH. Results are expressed as fold increase over untreated controls and are the mean of triplicate samples±SD. Statistical significance between IFN-γ alone samples and Pam₃CSK₄ prior to IFN-γ treated samples was determined by two-tailed t-test. Results are representative of at least two independent experiments.

Figure 5. TLR2 stimulation inhibits histone acetylation at the CIITA promoter, but not the CXCL11 promoter.

BMDM were treated with 10 ng/ml Pam₃CSK₄ for 8 h prior to stimulation with 20 ng/ml IFN-γ for 1, 2, 4, and 8 h. Cross-linked, sheared DNA was immunoprecipitated with antibodies against either acetylated histone H3 (A, B, C) or H4 (D, E, F). Precipitates were analyzed by qPCR using primers specific for the CXCL11 (A, D), CIITA (B, E), and NOS2 (C, F) promoters. Values were normalized to GAPDH and untreated controls and are the mean of triplicate samples±SD. *, p<0.01 comparing IFN-γ alone with Pam₃CSK₄ prior to IFN-γ treated samples (as determined by two-tailed t-test). Results are representative of two independent experiments.

Figure 6. IFN-γ-induced transcription of CXCL11, but not CIITA, is restored in TLR2 stimulated RelA^−/− macrophages.

BMDM from TNF^−/−/RelA^+/+ and TNF^−/−/RelA^−/− mice were treated with 10 ng/ml Pam₃CSK₄ for 8 h followed by 20 ng/ml IFN-γ for 4 h. Total RNA was harvested, reverse transcribed, and CXCL11 (A), CIITA (B), and NOS2 (C) expression analyzed by qPCR. All values were normalized to GAPDH and shown as fold induction compared to untreated samples without Pam₃CSK₄ or IFN-γ. Statistical significance was determined by two-tailed t-test between IFN-γ alone with Pam₃CSK₄ prior to IFN-γ samples. C57BL/6 BMDM showed similar results to those obtained with TNF^−/−/RelA^+/+ BMDM.