Inhibition of specific signaling pathways rather than epigenetic silencing of effector genes is the leading mechanism of innate tolerance

Abstract

Introduction: Macrophages activated through a pattern-recognition receptor (PRR) enter a transient state of tolerance characterized by diminished responsiveness to restimulation of the same receptor. Signaling-based and epigenetic mechanisms are invoked to explain this innate tolerance. However, these two groups of mechanisms should result in different outcomes. The epigenetic scenario (silencing of effector genes) predicts that activation of a PRR should broadly cross-tolerize to agonists of unrelated PRRs, whereas in the signaling-based scenario (inhibition of signaling pathways downstream of specific PRRs), cross-tolerization should occur only between agonists utilizing the same PRR and/or signaling pathway. Also, the so-called non-tolerizeable genes have been described, which acquire distinct epigenetic marks and increased responsiveness to rechallenge with the same agonist. The existence of such genes is well explained by epigenetic mechanisms but difficult to explain solely by signaling mechanisms.

Methods: To evaluate contribution of signaling and epigenetic mechanisms to innate tolerance, we tolerized human macrophages with agonists of TLR4 or NOD1 receptors, which signal via distinct pathways, and assessed responses of tolerized cells to homologous restimulation and to cross-stimulation using different signaling, metabolic and transcriptomic read-outs. We developed a transcriptomics-based approach to distinguish responses to secondary stimulation from continuing responses to primary stimulation.

Results: We found that macrophages tolerized with a NOD1 agonist lack responses to homologous restimulation, whereas LPS-tolerized macrophages partially retain the ability to activate NF-κB pathway upon LPS rechallenge, which allows to sustain low-level expression of a subset of pro-inflammatory genes. Contributing to LPS tolerance is blockade of signaling pathways required for IFN-β production, resulting in 'pseudo-tolerization' of IFN-regulated genes. Many genes in NOD1- or TLR4-tolerized macrophages are upregulated as the result of primary stimulation (due to continuing transcription and/or high mRNA stability), but do not respond to homologous restimulation. Hyperresponsiveness of genes to homologous rechallenge is a rare and inconsistent phenomenon. However, most genes that have become unresponsive to homologous stimuli show unchanged or elevated responses to agonists of PRRs signaling via distinct pathways.

Discussion: Thus, inhibition of specific signaling pathways rather than epigenetic silencing is the dominant mechanism of innate tolerance.

Keywords: NOD1; TLR4; innate immune response; lipopolysaccharide; macrophages; muramyl peptides; tolerance; transcriptome analysis.

Figure 1

Experimental design of the study and criteria of gene classification. (A) Design of cell culture experiments. (B) Criteria of gene classification based on ratios of RPM. Agonist 1 (A1) is the one used for the 1^st stimulation (24 h), agonist 2 (A2) for the 2^nd stimulation (1 or 4 h). Agonists 1 and 2 can be the same. 0, culture without agonists.

Figure 2

Responses of LPS-tolerized macrophages to LPS restimulation. Macrophages were cultured for 24 h with LPS (L) or medium alone (0), washed and recultured with LPS or medium alone for indicated periods of time. (A) Levels of TNF in supernatants of naïve and LPS-tolerized macrophages (ELISA, 6 experiments, M±σ). (B) Kinetics of ECAR of naïve and LPS-tolerized macrophages after addition of LPS (1 representative experiment out of 6, M±σ of quadruplicate wells). (C) Heat map of relative gene expression (RPM_sample/RPM_baseline) in RNA-seq experiment 1. (D–G) Statistics of relative gene expression (log₂-transformed) of T genes (D), all NT genes (E), SE genes (F), NH genes (G) in RNA-seq experiments 1 and 2. Boxes represent 25^th and 75^th percentiles, lines within boxes are medians, whiskers are 10^th and 90^th percentiles. P-values by repeated measures ANOVA with Tukey correction (*p < 0.05, ***p < 0.001, ****p < 0.0001, n.s. = non-significant).

Figure 3

Analysis of responses of LPS-tolerized macrophages to LPS restimulation. (A) Sizes of T, NT, SE and NH gene sets identified in RNA-seq experiments 1 and 2, and their intersections (numbers of genes). Shown in parentheses are expected numbers of coincidences if they were purely random. Differences between expected and actual numbers of coincidences were assessed using χ² test. (B) Gene sets significantly over-represented among the 118 NT genes (all with p < 0.001 in χ² test).

Figure 4

Relationship between mRNA expression patterns, mRNA transcription and mRNA stability in naïve and LPS-tolerized macrophages (RT-PCR). Cells were cultured 24 h without or with LPS, washed and recultured without or with LPS for indicated periods of time. (A, B) Kinetics of cytoplasmic mature mRNA expression and chromatin-associated RNA (caRNA) expression, respectively, in naïve and LPS-tolerized macrophages. One experiment out of two with similar results. >1, more than 1 log₂ difference between LPS➔0 and LPS➔LPS macrophages. Log₂ mRNA or caRNA expression in untreated macrophages equals zero. (C) stability of mature cytoplasmic mRNA in naïve macrophages stimulated by LPS and in LPS-tolerized macrophages recultured with or without LPS (3 experiments, bars denote means). Actinomycin D (ActD) was added at 1 h of (re)stimulation and cells were further cultured for 1 h. Results are expressed as ratios of mRNA levels after and before addition of ActD.

Figure 5

Responses of M-triDAP-tolerized macrophages to M-triDAP restimulation. Macrophages were cultured for 24 h with M-triDAP (M) or medium alone (0), washed and recultured without or with M-triDAP for indicated periods of time. (A) Levels of TNF in supernatants of naïve and M-triDAP-tolerized macrophages after restimulation with M-triDAP (ELISA, 6 experiments, M±σ). (B) Kinetics of ECAR of naïve and M-triDAP-tolerized macrophages after addition of M-triDAP (1 representative experiment out of 6, M±σ of quadruplicate wells). (C) Heat map of relative gene expression (RPM_sample/RPM_baseline) in RNA-seq experiment 1. (D, E) Statistics of relative gene expression (log₂-transformed) of T genes (D) and all NT genes (E) in RNA-seq experiments 1 and 2. (F) Sizes of T, NT, SE and NH gene sets identified in RNA-seq experiments 1 and 2, and their intersections (numbers of genes). Shown in parentheses are expected numbers of coincidences if they were purely random. Statistical analysis as in Figures 2 and 3 . *p < 0.05, ***p < 0.001, ****p < 0.0001, n.s. = non-significant.

Figure 6

Activation of NF-κB and p38 pathways in naïve, M-triDAP- and LPS-tolerized macrophages after their reculture with medium (0), M-triDAP (M) or LPS (L). (A) Levels of TNF and IL-6 in macrophage supernatants after indicated stimulation sequences (3 to 9 experiments). Statistical analysis by repeated measures ANOVA with Tukey correction. * p < 0.05, n.s. = non-significant. (B) Immunoblotting for IκBα, phospho-p38, phospho-ERK1/2, phospho-JNK, total p38, total ERK1/2 and α-tubulin at indicated time points after 2^nd stimulation (one representative Western blot out of three) and corresponding densitometry data normalized to untreated macrophages at 15 min. (C) Effect of IκB kinase inhibitor on TNF, IL6 and IL1B mRNA expression in naïve and LPS-tolerized macrophages (RT-PCR). Cells were cultured without or with LPS for 24 h, washed, incubated without or with PF 184 (5 μM) for 15 min, whereafter recultured without or with LPS for 1 or 4 h. 3 experiments, bars denote means. *p < 0.05, **p < 0.01 for comparisons between cells treated or not with PF 184 in identical stimulation conditions (paired t-test).

Figure 7

Responses of macrophages tolerized with M-triDAP (M) to stimulation with LPS (L). Only genes inducible in naïve macrophages both by M-triDAP and by LPS in both experiments after 1 and/or 4 h of stimulation are included (n = 446). (A) General statistics of T and NT gene expression in RNA-seq experiments 1 and 2 upon M-triDAP➔LPS stimulation. Responses of naïve macrophages to LPS (0➔LPS), of M-triDAP-tolerized macrophages to M-triDAP (M-triDAP➔M-triDAP) as well as residual M-triDAP-induced expression (M-triDAP➔0) are shown in parallel. (B) Heat map of relative gene expression from RNA-seq experiment 1. Statistical analysis as in Figure 2 . **p < 0.01, ****p < 0.0001, n.s. = non-significant.

Figure 8

Intersections of T and NT gene sets identified in RNA-seq experiments. Only 446 genes inducible in naïve macrophages both by LPS and by M-triDAP in both RNA-seq experiments after 1 and/or 4 h of stimulation are included. (A) Sizes of T and NT gene sets identified in RNA-seq experiment 1 and 2 upon different stimulation sequences, and their intersections. (B) Venn diagrams showing intersections of T and NT gene sets upon M-triDAP➔M-triDAP and M-triDAP➔LPS stimulation. (C) Venn diagrams showing intersections of T and NT gene sets upon LPS➔LPS and LP➔M-triDAP stimulation.

Figure 9

Responses of macrophages tolerized with LPS (L) to stimulation with M-triDAP (M). Only genes inducible in naïve macrophages both by M-triDAP and by LPS in both experiments after 1 and/or 4 h of stimulation are included (n = 446). (A) General statistics of T and NT gene expression in RNA-seq experiments 1 and 2 upon LPS➔M-triDAP stimulation. Responses of naïve macrophages to M-triDAP (0➔M-triDAP), of LPS-tolerized macrophages to LPS (LPS➔LPS) as well as residual LPS-induced gene expression (LPS➔0) are shown in parallel. (B) Heat map of relative gene expression from RNA-seq experiment 1. Statistical analysis as in Figure 2 . *p < 0.05, ****p < 0.0001, n.s. = non-significant.

Figure 10

Regulation of T and NT gene expression in naïve and LPS-tolerized macrophages. (A) Numbers of T and NT genes containing indicated TFBS in their promoters. Only genes showing identical behavior in two RNA-seq experiments (see Figure 8A ) are included in the analysis. Shown TFs satisfy following criteria: (i) TFBS are present in at least 10% of T or NT genes; (ii) enrichment score (ES) > 1; (iii) significant difference between frequencies of TFBS in T and NT gene sets (p < 0.01 by χ²-test). Numbers to the right of the plots indicate ES. (B) Representation of genes from indicated gene sets among T and NT genes upon different stimulation sequences. (C, D) Relative expression of IFNB1 and MX1 mRNA, respectively, after different stimulation sequences at 1 and 4 h of 2^nd stimulation. Horizontal bars denote means. (E) Effect of exogenous IFN-β (I) on MX1 mRNA expression in naïve and LPS-tolerized macrophages. Cells were cultured for 24 h with medium (0) or with LPS (L, 100 ng/ml), then recultured for 1 or 4 h with medium, LPS (100 ng/ml) or rhIFN-β1b (10 U/ml). (F, G) Effect of liposomal poly-I:C (P) on IFNB1 and MX1 mRNA expression, respectively, in naïve and LPS-tolerized macrophages. Naïve or LPS-tolerized macrophages were recultured for 1 or 4 h with medium (0) or liposomal poly-I:C (1 μg/ml). (H, I) Expression of E2F1 and E2F2 mRNA in naïve and LPS-tolerized macrophages recultured with medium (0) or LPS (L) for 1 or 4 h. In (C, D, H, I), red and black symbols denote RNA-seq and RT-PCR data, respectively. *p < 0.05, ***p < 0.001, n.s. = non-significant.