IRAK4 kinase activity is not required for induction of endotoxin tolerance but contributes to TLR2-mediated tolerance

IRAK4 kinase-inactive macrophages exhibit attenuated TLR2- and TLR4-mediated signaling and normal induction of endotoxin tolerance, while TLR2-mediated homotolerance, and TLR2-mediated TLR4 heterotolerance, are deficient.

Abstract

Prior exposure to LPS induces “endotoxin tolerance” that reprograms TLR4 responses to subsequent LPS challenge by altering expression of inflammatory mediators. Endotoxin tolerance is thought to limit the excessive cytokine storm and prevent tissue damage during sepsis but renders the host immunocompromised and susceptible to secondary infections. Tolerance initiated via one TLR can affect cellular responses to challenge via the same TLR (“homotolerance”) or through different TLRs (“heterotolerance”). IRAK4, an essential component of the MyD88-dependent pathway, functions as a kinase and an adapter, activating subsets of divergent signaling pathways. In this study, we addressed mechanistically the role of IRAK4 kinase activity in TLR4- and TLR2-induced tolerance using macrophages from WT versus IRAK4^KDKI mice. Whereas IRAK4 kinase deficiency decreased LPS signaling, it did not prevent endotoxin tolerance, as endotoxin pretreatment of WT and IRAK4^KDKI macrophages inhibited LPS-induced MAPK phosphorylation, degradation of IκB-α and recruitment of p65 to the TNF-α promoter, expression of proinflammatory cytokines, and increased levels of A20 and IRAK-M. Pretreatment of WT macrophages with Pam3Cys, a TLR2–TLR1 agonist, ablated p-p38 and p-JNK in response to challenge with Pam3Cys and LPS, whereas IRAK4^KDKI macrophages exhibited attenuated TLR2-elicited homo- and heterotolerance at the level of MAPK activation. Thus, IRAK4 kinase activity is not required for the induction of endotoxin tolerance but contributes significantly to TLR2-elicited homo- and heterotolerance.

Introduction

Sensing of invading microbial pathogens by PRRs triggers expression of cytokines and antimicrobial effectors by neutrophils, macrophages, and DCs, up-regulates expression of MHC and costimulatory molecules, and primes adaptive immune responses [1–3]. TLRs represent one important class of PRRs expressed on the cell membrane (e.g., TLR1, TLR2, TLR4, TLR5, TLR6) or intracellular endosomes (TLR3, TLR7–9) of immune, epithelial, and endothelial cells [4]. TLR ectodomains recognize conserved microbe-associated molecular patterns (e.g., LPS), initiating TLR dimerization that brings together intracellular Toll-IL-1R resistance domains, providing docking platforms that facilitate recruitment of adapter proteins and kinases [1, 5]. Kinase activation initiates downstream signaling cascades that culminate in expression of inflammatory mediators, MHC antigens, adhesion, and costimulatory molecules [3]. TLR4 signals from the cell surface and early endosomes upon recruitment of MyD88 [3] to trigger expression of proinflammatory cytokines and chemokines [6, 7]. Translocation of TLR4 and the TRIF-related adapter molecule into endolysosomal compartments allows recruitment of signaling adapters TRIF and TRAF3, leading to activation of IFN regulatory factor 3 and expression of IFN-β [8–10].

Upon activation of the MyD88-dependent pathway, TLR4-associated MyD88 forms “Myddosome” complexes [11] by recruiting IRAK4, -2, and -1 kinases [12]. Clustering of IRAK4 triggers kinase activation, resulting in IRAK4-mediated phosphorylation and activation of IRAK1 and IRAK2, their association with TRAF6, and engagement of TAK1 [1, 3, 12, 13]. TAK1 subsequently activates MAPKs and the classical pathway of NF-κB activation by activating IKK-β, p-IκB and IκB proteasomal degradation, NF-κB translocation to the nucleus, and transcription of NF-κB-dependent inflammatory genes [14–16]. Kinase-deficient IRAK4 functions as an adapter by engaging MEKK3, leading to p-IKK-γ, activation of IKK-α, p-IκB and IκB dissociation from NF-κB subunits without IκB-α proteasomal degradation, and expression of cytokine genes [13, 16]. Differential dependence of various cytokine genes on IRAK4 kinase activity versus its adapter function [14, 15, 17] is thought to underlie complex fine-tuning mechanisms of regulation of TLR signaling during infection and inflammation.

Development of endotoxin tolerance following the initial “cytokine storm” phase of sepsis is thought to protect the host from an overexuberant immune response and tissue damage but at the same time, may render the host immunocompromised and more susceptible to secondary infection [18–20]. “Reprogramming” [21] of TLR4 signaling in endotoxin-tolerant monocytes and macrophages does not occur as a result of decreased TLR4 expression but involves altered recruitment, tyrosine phosphorylation, and K63-linked polyubiquitination of proximal receptor-adapter-kinase complexes [22–27] and induction of negative regulators IRAK-M, SHIP1, and A20 [24, 25, 28]. Although a few studies have sought to dissociate kinase and adapter functions of IRAK4 in IL-1R/TLR signaling, albeit with conflicting results [13–16, 29–31], it is unclear how IRAK4 kinase activity affects induction of TLR2 and TLR4 homo- and heterotolerance. To address these questions, we used IRAK4^KDKI mice to determine the impact of kinase deficiency of IRAK4 on the induction of TLR tolerance. Our data showed comparable induction of endotoxin tolerance in WT or IRAK4^KDKI PMs and BMDMs, as judged by attenuated MAPK phosphorylation, inhibited expression of proinflammatory cytokines and chemokines, and up-regulation of negative TLR regulators, A20 and IRAK-M. Notably, IRAK4 kinase activity was found to be a prerequisite for conferring inhibition of LPS-inducible JNK and p38 MAPK activation following prior exposure to Pam3Cys. These results represent the first systematic analyses of the role of IRAK4 kinase activity in TLR homo- and heterotolerance and pave the way for improved understanding of how IRAK4 kinase dysregulation may underlie immunocompromised states in late sepsis.

MATERIALS AND METHODS

Reagents

Highly purified, protein-free Escherichia coli K235 LPS was prepared as described [32], and Pam3Cys was purchased from Invivogen (San Diego, CA, USA). The following antibodies were used in this study: anti-p- and anti-total p38, anti-p- and anti-total ERK, anti-p- and anti-total JNK (Promega, Madison, WI, USA), anti-p- and anti-IκB-α, anti-IRAK1, anti-IRAK-M, anti-A20, anti-p65, and anti-tubulin (Santa Cruz Biotechnology, Santa Cruz, CA, USA).

Mice, macrophage isolation, and cell culture

C57BL/6J mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). Female IRAK4^KDKI were derived on a C57BL/6 background [15] and were generously provided by Drs. Kirk A. Staschke and Raymond Gilmour (Lilly Research Laboratories, Indianapolis, IN, USA), through a Materials Transfer Agreement, and were bred homozygously at the University of Maryland, Baltimore (Baltimore, MD, USA). Four- to 6-week-old female mice were used in experiments. All of the animal protocols used in this study were carried out with institutional approval. Peritoneal exudate cells were isolated from mice 4 days after i.p. injection with sterile 3% thioglycollate (Remel, Lenexa, KS, USA), and PMs were obtained by peritoneal lavage and subsequent adherence to plastic, as described [28, 33]. Cells were plated into six-well plates (4×10⁶ cells/well) and cultured in RPMI 1640 (Mediatech, Herndon, VA, USA), supplemented with 10% FBS (HyClone, Logan, UT, USA), 2 mM L-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin (cRPMI; Mediatech). BM cells were isolated from femurs and tibias of mice and were cultured in cRPMI, supplemented with 25% L929 cell-conditioned supernatant that was used as a source of M-CSF to generate BMDMs, as reported previously [34]. Cells were pretreated for 20 h with culture medium (controls), 10 ng/ml LPS, or 100 ng/ml Pam3Cys, washed, and restimulated, as indicated in the figure legends. All animal procedures were carried out with Institutional Animal Care and Use Committee approval.

RNA isolation and qRT-PCR analysis

Total RNA was isolated with Trizol (Invitrogen, Carlsbad, CA, USA), treated with DNase, and repurified, as recommended by the manufacturer. cDNA was prepared from 1 μg total RNA, using a reverse transcription system (Promega), and examined by qRT-PCR, using 5 μl cDNA, 0.3 μM gene-specific primers, and SYBR Green Supermix (Bio-Rad, Hercules, CA, USA) on a MyiQ real-time PCR machine (Bio-Rad). The following primers were used: mouse hypoxanthine-guanine phosphoribosyl transferase: forward, 5′-GCTGACCTGCTGGATTACATT-3′, reverse, 5′-GTTGAGAGATCATCTCCACCA-3′; TNF-α: forward, 5′-ACCCTCACACTCAGATCATCT-3′, reverse, 5′-TTGTCTTTGAGATCCATGCCGT-3′; IL-6: forward, 5′-TCAGGAAATTTGCCTATTGAAA ATTT-3′, reverse, 5′-GCTTTGTCTTTCTTGTTATCTTTTAAGTTGT-3′; KC: forward, 5′-TGTCAGTGCCTGCAGACCAT-3′, reverse, 5′-GCTATGACTTCGGTTTGG GTG-3′; pro-IL-1β: forward, 5′-ACAGAATATCAACCAA CAAGTGATATTCTC-3′, reverse, 5′-GATTCTTTCCTTTGAGGCCCA-3′; IL-12p40: forward, 5′-TCTTTGTTCGAATCCAGC GC-3′, reverse, 5′-GGAACGCACCTTTCTGGT TACA-3′; RANTES: forward, 5′-GAGTGACAAACACGACTGCAAGAT-3′, reverse, 5′-CTGCTTTGCCTACCTCTCCC T-3′; IFN-β: forward, 5′-CCCTATGGAGATGACGGAGA-3′, reverse, 5′-GTCTCATTCCACCCAGTGCT-3′; IRAK-M: forward, 5′-CATCAACTATGGAGTAAGCTGGAC-3′, reverse, 5′-GTCCAGGGTCGTTTTCTCTG-3′; A20: forward, 5′-TACGACACTCGGAACTGGAAT-3′, reverse, 5′-TGACAATGATGGGTCTTCTGA-3′. Data were analyzed by the 2^{−ΔΔcomparative threshold} method [35] and presented as fold-induction, normalized to cells exposed to medium only.

ChIP assay

ChIP assays were performed as described [36]. Cells were fixed in 1% formaldehyde solution and washed with ice-cold PBS, and a glycine stop solution was added for 5 min. After removing the stop solution and addition of ice-cold PBS containing PMSF, cells were scrapped off and centrifuged, and the pellet was resuspended in a lysis buffer containing protease inhibitor cocktail and incubated for 30 min. Enzymatic shearing cocktail was added, and samples were homogenized in dounce homogenizers and centrifuged. The nuclei were resuspended in a digestion buffer with the enzymatic shearing cocktail and incubated for 15 min. The sheared chromatin was processed using the ChIP-IT Express kit (Active Motif, Carlsbad, CA, USA) and immunoprecipitated with anti-p65 or isotype control antibodies. The samples were analyzed by PCR using primer pairs that span the p65-binding sites in the promoter regions of mouse TNF-α, IFN-β, and RANTES as follows: TNF-α promoter: forward, 5′-TCCTTGATGCCTGGGTGTCCC-3′, reverse, 5′-GCAGACGGCCGCCTTTATAGC-3′; IFN-β promoter: forward, 5′-CTGTCAAAGGCTGCAGTGAG-3′; reverse, 5′-GCCAGGAGCTTGAATAAAATG-3′; and RANTES promoter: forward, 5′-TGACACAAGTGTGG TCTGTTTCTG-3′, reverse, 5′-AGGTAGCAGGGAGCTGTTGTCTTA-3′.

Preparation of cell extracts and Western blot analyses

Cells were lysed for 30 min in an ice-cold buffer containing 50 mM Tris-HCl (pH 7.4), 1 mM PMSF, 1 mM DTT, 1 mM sodium orthovanadate, 50 mM NaF, 2 mM EDTA, 150 mM NaCl, 1% Triton X-100, 20 mM β-glycerol phosphate, and protease inhibitor cocktail (Roche Applied Science, Indianapolis, IN, USA). After centrifugation (12,000 g, 4°C, 20 min), supernatants were collected and protein concentration determined using the DC Protein Assay kit (Bio-Rad). Cell lysates were resuspended in Laemmli sample buffer (Bio-Rad) and boiled for 10 min. Proteins (20 μg) were separated on 4–10% polyacrylamide gels (Invitrogen) and electrotransferred to Immobilon-P membranes (Millipore, Bedford, MA, USA). Membranes were blocked and probed with the indicated antibodies, and bands were visualized by ECL, as described [22, 24, 25, 28]. Band intensities were quantified using Quantity One software (Bio-Rad).

Analysis of cytokine production

Levels of cytokines in cell-free supernatants were determined by ELISA, using antibodies and standards from R&D Systems (Minneapolis, MN, USA). The lower detection limits were 3.9, 6.3, 3.1, and 9.3 pg/ml for TNF-α, IL-1β, KC, and RANTES, respectively.

Statistical analysis

Data were analyzed with the GraphPad Prism 5 program for Windows (GraphPad Software, San Diego, CA, USA), using a one-way ANOVA with repeated measures, followed by post hoc comparisons with Tukey's multiple paired comparison test. The results were expressed as mean ± sd values.

RESULTS

Comparable suppression of LPS-inducible proinflammatory cytokines and chemokines in endotoxin-tolerant WT and IRAK4^KDKI macrophages

To define the role of IRAK4 kinase activity in endotoxin tolerance, we pretreated WT and IRAK4^KDKI macrophages for 20 h with medium (control) or exposed to LPS (endotoxin-tolerized group) and examined expression of proinflammatory cytokine genes in response to subsequent LPS challenge. Control WT PMs and BMDMs exhibited robust LPS-inducible cytokine mRNA expression that was consistently greater than that seen in medium-pretreated IRAK4^KDKI macrophages (Figs. 1 and 2). The extent to which deficiency in IRAK4 kinase activity affects LPS inducibilty varied among cytokine genes. In PMs, the LPS-mediated expression of pro-IL-1β mRNA was the most significantly curtailed in IRAK4^KDKI cells (72–93% inhibition vs. levels observed in WT cells), whereas the genes encoding TNF-α, KC, and IL-12 p40 were less affected (34–67% suppression; Fig. 1). The LPS inducibilty of these cytokine genes was also inhibited by 40–70% in IRAK4^KDKI BMDMs initially exposed to medium (Fig. 2A–C). Endotoxin-pretreated WT PMs were tolerized, as evidenced by a 45–85% inhibition of LPS-inducible TNF-α mRNA (Fig. 1A) and 88–96% suppression of pro-IL-1β, KC, and IL-12 p40 mRNA (Fig. 1B–D). Despite the partial loss of LPS responsiveness, IRAK4^KDKI PMs (Fig. 1) and BMDMs (Fig. 2) retained the capacity to develop endotoxin tolerance, as judged by 57–95% inhibition of LPS-induced expression of these same cytokine genes.

Figure 1. Induction of endotoxin tolerance in WT and IRAK4^KDKI PMs inhibits LPS-mediated expression of proinflammatory cytokine genes.

Thioglycollate-elicited PMs from WT and IRAK4^KDKI mice were pretreated for 20 h with medium or LPS (10 ng/ml), washed three times, and exposed to medium or restimulated with 100 ng/ml LPS for the indicated time periods. RNA was reverse-transcribed and analyzed by qRT-PCR with the respective gene-specific primers. The summary of three independent experiments (A and C) and the data of representative experiments (n=3; B and D) are shown. *P < 0.05.

Figure 2. Induction of endotoxin tolerance in WT and IRAK4^KDKI BMDMs inhibits LPS-mediated expression of proinflammatory cytokine genes.

WT and IRAK4^KDKI BMDMs were treated as described in the legend to Fig. 1. RNA was reverse-transcribed and examined by qRT-PCR. The results of representative experiments (n=3) are shown. *P < 0.05.

To confirm and extend these results, we next examined the impact of IRAK4 kinase deficiency on LPS-mediated secretion of proinflammatory cytokines in control (medium-pretreated) and LPS-pretreated macrophages. LPS challenge, 20 h later, led to robust secretion of TNF-α, KC, and IL-12 p70 in medium-pretreated WT PMs and BMDMs, whereas LPS inducibilty of these cytokines in IRAK4^KDKI control cells was inhibited by 50–75% (Table 1). Prior endotoxin exposure of WT and IRAK4^KDKI macrophages attenuated LPS-inducible cytokine secretion by 50–96% compared with the cytokine responses of medium-pretreated cells (Table 1). These data indicate that IRAK4 kinase deficiency does not prevent induction of endotoxin tolerance.

Table 1. Cytokine Secretion in Control and LPS-Tolerant WT and IRAK4^KDKI Macrophages.

First pretreat/second challenge	WT PMs	IRAK4KDKI PMs	WT BMDMs	IRAK4KDKI BMDMs
TNF-α (pg/ml)
Medium/medium	11.8 ± 2.5	10.3 ± 3.5	11.4 ± 3.5	13.0 ± 4.6
Medium/LPSa	906.3 ± 74.0	562.1 ± 135.4b	675.5 ± 18.6	393.8 ± 88.4b
LPS/medium	10.2 ± 3.2	14.0 ± 2.8	10.0 ± 4.4	12.8 ± 3.5
LPS/LPSc	294.2 ± 95.6b	226.4 ± 75.0b	124.0 ± 30.6b	48.5 ± 10.6b
KC (pg/ml)
Medium/medium	63.9 ± 17.4	32.4 ± 2.8	36.9 ± 0.6	25.5 ± 1.1
Medium/LPSa	7176.0 ± 369.7	5351.1 ± 135.4b	4079.2 ± 174.1	2424.3 ± 275.3b
LPS/medium	165.0 ± 36.0	165.9 ± 43.1	235.9 ± 8.6	39.8 ± 0.7
LPS/LPSc	2500.0 ± 657.5b	1306.5 ± 230.2b	1946.8 ± 99.2b	222.2 ± 5.1b
IL-12 p70 (pg/ml)
Medium/medium	18.2 ± 0.1	25.8 ± 3.0	12.3 ± 5.7	12.3 ± 5.6
Medium/LPS	1088.1 ± 100.6	673.1 ± 35.2b	138.0 ± 31.5	111.7 ± 37.2
LPS/medium	22.7 ± 0.1	26.4 ± 3.5	12.7 ± 5.2	22.3 ± 11.1
LPS/LPS	52.8 ± 3.1b	27.1 ± 1.1b	14.9 ± 3.4b	15.3 ± 2.8b
RANTES (pg/ml)
Medium/medium	19.8 ± 11.1	19.9 ± 12.0	16.3 ± 5.4	18.0 ± 5.9
Medium/LPS	1687.1 ± 313.3	1093.4 ± 239.9	4070.1 ± 1649.0	2827.3 ± 470.8
LPS/medium	1669.1 ± 287.4	3166.2 ± 340.3	1175.3 ± 327.7	770.4 ± 80.9
LPS/LPS	3571.4 ± 843.5	3898.5 ± 803.5	10,020.4 ± 4216.1	4520.1 ± 1871.2

Macrophages were pretreated as described in the legend to Fig. 1. Twenty-four hours post restimulation with medium or 100 ng/ml LPS, supernatants were analyzed by ELISA.

^aDifferences between cytokine levels secreted by medium-pretreated, LPS-restimulated WT versus IRAK4^KDKI macrophages were compared.

^bP < 0.05.

^cThe data in the medium-pretreated, LPS-restimulated groups were compared with the results in the LPS-pretreated, LPS-restimulated groups. The results (mean±sd) of three independent experiments are shown.

Induction of endotoxin tolerance in WT and IRAK4^KDKI macrophages similarly inhibits LPS-mediated expression of IFN-β mRNA but does not impact RANTES

Contradictory data have been reported with respect to the involvement of IRAK4 in regulating the expression of TRIF-dependent genes [30, 37–39]. Therefore, we next examined the role of IRAK4 kinase activity in LPS inducibilty and endotoxin tolerance of TRIF-dependent cytokine genes. LPS challenge of medium-pretreated WT and IRAK4^KDKI PMs (Fig. 3) and BMDMs(data not shown) resulted in similar induction patterns of IFN-β and RANTES mRNA, except for a delay in the induction of RANTES mRNA at 2 h post-LPS treatment of IRAK4^KDKI cells and comparable increases in secretion of these cytokines (Table 1). Similar to previously reported results in WT macrophages [40, 41], LPS pretreatment of WT and IRAK4^KDKI cells resulted in marked down-regulation of LPS-induced IFN-β mRNA (Fig. 3A), whereas expression of RANTES mRNA and protein was unaffected (Fig. 3B and Table 1). Whereas IRAK4^KDKI macrophages showed lower LPS-inducible TNF-α mRNA expression compared with WT cells, they exhibited comparable responses to poly (I:C) (Fig. 4), an agonist that activates TLR3 and uses TRIF exclusively [3, 42]. Prior exposure to LPS attenuated up-regulation of TNF-α mRNA induced by subsequent LPS challenge, whereas not affecting responses to poly (I:C) (Fig. 4), consistent with previous reported results [43]. Thus, IRAK4 kinase deficiency does not affect the LPS inducibilty of the TRIF-dependent cytokines, IFN-β and RANTES, does not impact poly (I:C)-mediated, TRIF-dependent induction of TNF-α mRNA, and does not affect the ability to tolerize IFN-β gene expression. These results support the selective, specific impact of IRAK4 kinase deficiency on the MyD88-signaling pathway and demonstrate the lack of off-target effects of knocked-in, kinase-deficient IRAK4 species. Furthermore, our data showing differential sensitivity of IFN-β and RANTES to endotoxin tolerance suggest distinct mechanisms of reprogramming of these cytokine genes.

Figure 3. Endotoxin pretreatment of WT and IRAK4^KDKI PMs inhibits LPS-mediated expression of IFN-β but does not affect levels of RANTES mRNA.

PMs from WT and IRAK4^KDKI mice were treated, as described in the legend to Fig. 1. RNA samples were subjected to reverse transcription and qRT-PCR with the corresponding gene-specific primers. The results of representative experiments (n=3) are depicted. *P < 0.05.

Figure 4. The impact of IRAK4 kinase deficiency on LPS- and poly (I:C)-mediated induction of TNF-α gene expression in control and LPS-tolerized macrophages (Mϕs).

WT and IRAK4^KDKI PMs were exposed for 20 h to medium or 10 ng/ml LPS, washed, and challenged for 3 h with medium, 100 ng/ml LPS, or 50 μg/ml poly (I:C). RNA was isolated and analyzed by real-time qPCR with the respective gene-specific primers. Shown are the data of representative (n=3) experiments. *P < 0.05.

Decreased LPS-mediated p-ERK, p-JNK, and p-p38, activation of NF-κB, and increased expression of IRAK-M and A20 in endotoxin-tolerant WT and IRAK4^KDKI macrophages

Next, the impact of endotoxin tolerance induction on activation of intracellular signaling intermediates in WT and IRAK4^KDKI PMs was examined. Endotoxin-tolerant monocytes and macrophages exhibit suppressed LPS-mediated activation of IRAK1, MAPKs, and NF-κB [22–24, 33, 44] and up-regulated expression of IRAK-M and A20, negative regulators of TLR signaling [25, 28, 45–48]. Following LPS stimulation, IRAK1 undergoes phosphorylation and ubiquitination [49–53], resulting in appearance of higher molecular-weight IRAK1 species and disappearance of unmodified IRAK1 bands [24, 25, 49]. LPS challenge of medium-pretreated WT PMs resulted in disappearance of unmodified IRAK1 and degradation of IκB-α, with a concomitant increase in p-ERK, p-JNK, and p-p38 MAPKs (Fig. 5A and B). No significant alterations in the levels of unmodified IRAK1 were detected in LPS-stimulated, medium-pretreated IRAK4^KDKI PMs, whereas LPS-inducible degradation of IκB-α and p-ERK, p-JNK, and p-p38 were attenuated (Fig. 5). LPS pretreatment of WT or IRAK4^KDKI PMs suppressed degradation of IκB-α and p-ERK, p-JNK, and p-p38 in response to subsequent LPS challenge (Fig. 5). ChIP assays demonstrated deficient LPS-induced recruitment of the p65 NF-κB subunit to the TNF-α promoter in IRAK4^KDKI versus WT PMs but comparable recruitment of p65 to the IFN-β and RANTES promoters (Fig. 6). Upon prior exposure to LPS, WT and IRAK4^KDKI PMs exhibited similar decreases in LPS-mediated p65 recruitment to TNF-α and IFN-β promoters compared with medium-pretreated cells, documenting endotoxin tolerance (Fig. 6). Induction of endotoxin tolerance in WT and IRAK4^KDKI PMs led to five- to 6.2-fold increases in A20 mRNA and an approximate twofold increase in IRAK-M mRNA (Fig. 7A; mRNA levels in LPS-pretreated PMs were normalized to those detected in medium-pretreated cells). LPS-tolerized WT and IRAK4^KDKI PMs exhibited comparably up-regulated levels of IRAK-M protein, whereas A20 protein was increased to a greater extent in WT cells (Fig. 7B). Thus, IRAK4 kinase activity is required for LPS-mediated changes in unmodified IRAK1, significantly contributes to TLR4-driven MAPK phosphorylation and NF-κB activation, but is dispensable for induction of endotoxin tolerance.

Figure 5. IRAK4^KDKI PMs show attenuated LPS-induced changes in unmodified IRAK1, degradation of IκB-α, and p-p38, but retain the capacity to develop endotoxin tolerance.

WT and IRAK4^KDKI PMs were pretreated for 20 h with medium or 10 ng/ml LPS, washed, and restimulated with medium or 100 ng/ml LPS for the indicated times. Cell lysates were subjected to SDS-PAGE and Western blot with antibodies against IRAK1 (A); IκB-α, p-p38, and tubulin (A and B); and p-ERK and p-JNK (B). The results of one of the three independent experiments are shown.

Figure 6. Recruitment of p65 to cytokine gene promoters in control and LPS-tolerized WT and IRAK4^KDKI macrophages.

After exposure for 20 h to medium or 10 ng/ml LPS, WT and IRAK4^KDKI PMs were washed and treated for 30 min with medium or 100 ng/ml LPS. Chromatin was isolated, sheared, immunoprecipitated (IP) with anti-p65 or isotype control antibodies, and processed for ChIP assays, as described in Materials and Methods (A). (B) Quantification of the results shown in A. Band intensities in immunoprecipitated samples were normalized for those obtained for total input samples and expressed as arbitrary units, reflecting p65 promoter recruitment. Shown are the data of representative (n=3) experiments.

Figure 7. Effect of endotoxin tolerance on expression of A20 and IRAK-M in WT and IRAK4^KDKI macrophages.

WT and IRAK4^KDKI PMs were treated for 20 h with medium or 10 ng/ml LPS, followed by preparation of RNA (A) or cell lysates (B). (A) RNA was subjected to reverse transcription and qRT-PCR. The results of one of the three independent experiments are presented. (B) Cell lysates were analyzed by Western blotting with the indicated antibodies. The results of representative experiments (n=3) are shown.

IRAK4 kinase deficiency impairs induction of TLR2 homotolerance and ablates Pam3Cys-mediated TLR4 heterotolerance

Next, we studied the impact of IRAK4 kinase deficiency on TLR2 signaling, TLR2 homotolerance (altered responses to Pam3Cys, a TLR2 agonist, following exposure to the same agonist), and TLR2-mediated heterotolerance of TLR4 (altered responses to LPS following exposure to Pam3Cys). TLR-inducible p-JNK and p-p38 MAPKs was examined, as it correlates with kinase activation [54] and has been widely used in studies on TLR signaling and tolerance [22–24, 33, 44, 46, 55]. Pam3Cys, a TLR2-TLR1 agonist [56, 57], induced robust p-JNK and p-p38 in medium-pretreated WT PMs at 15 and 180 min postchallenge, whereas attenuated MAPK phosphorylation was seen IRAK4^KDKI PMs (Fig. 8A). Prior exposure to Pam3Cys mitigated the capacity of WT PMs to respond to subsequent challenge with Pam3Cys, leading to 75% and 99% inhibition of TLR2-inducible p-p38 and p-JNK at 15 min postchallenge (Fig. 8A). Pam3Cys-pretreated IRAK4^KDKI PMs showed 41% and 17% suppression of the Pam3Cys-inducible levels of p-JNK and p-p38 at 15 min, compared with 90% and 99% inhibition of these responses exhibited by WT PMs. These data indicate that IRAK4 kinase deficiency impairs the induction of TLR2 homotolerance.

Figure 8. The effect of pretreatment of WT and IRAK4^KDKI PMs with Pam3Cys on Pam3Cys- and LPS-mediated p-JNK and p-p38 MAPKs.

WT and IRAK4^KDKI PMs were exposed for 20 h to medium or 1 μg/ml Pam3Cys, washed, treated with medium, or restimulated with 1 μg/ml Pam3Cys (A; homotolerance) or 100 ng/ml LPS (B; heterotolerance) for the indicated times. Cell lysates were analyzed by SDS-PAGE and Western blotting with the indicated antibodies. Right panels show quantification (arbitrary values) of the p-p38 and p-JNK bands normalized to tubulin levels. The results of representative (n=2) experiments are shown.

Figure 8B shows a similar experiment in which medium- or Pam3Cys-pretreated WT or IRAK4^KDKI PMs were challenged with LPS (TLR2-mediated TLR4 heterotolerance). As seen for LPS-induced cytokine gene expression, medium-pretreated WT PMs showed robust p-JNK in response to LPS, whereas the response of IRAK4^KDKI PMs to LPS was blunted. In Pam3Cys-pretreated WT PMs, challenge with LPS resulted in heterotolerance, as evidenced by 51% and 90% inhibition in LPS-inducible p-p38 and p-JNK MAPKs. In contrast, IRAK4^KDKI macrophages failed to respond to Pam3Cys pretreatment by inhibiting p-p38 upon subsequent challenge with LPS (Fig. 8B) and showed a more modest reduction of Pam3Cys- and LPS-inducible p-JNK. These results suggest that IRAK4 kinase activity plays an important role in eliciting TLR2-mediated MAPK activation, and cells expressing kinase-inactive IRAK4 have deficient capacities to develop TLR2 homotolerance or TLR2-elicited TLR4 heterotolerance.

DISCUSSION

We [24–26, 28, 33, 40, 41] and others [18, 19, 44–47] have shown reprogramming of TLR4-driven MyD88- and TRIF-dependent cytokine genes, the failure of LPS-inducible activation of IRAK4, and increased expression of negative regulators of TLR signaling in endotoxin-tolerant human monocytes and mouse macrophages. However, the role of IRAK4 kinase activity in TLR2 and TLR4 homotolerance and TLR2-mediated heterotolerance of TLR4 has not been explored. Furthermore, conflicting results have been published with respect to how IRAK4 kinase deficiency affects TLR2-, TLR7-, and TLR4-elicited signaling [14, 15, 17, 29–31, 39]. This paper demonstrates that IRAK4 kinase deficiency markedly decreases LPS-mediated changes in the levels of pro-IL-1β mRNA, unmodified IRAK1, and TLR2-mediated p-JNK and p-p38 but exerts significantly lower impact on TLR4-elicited MAPK activation and expression of TNF-α, KC, and IL-12. TLR engagement activates the kinase activity of IRAK4 following assembly of the Myddosome complex, recruitment and clustering of IRAK4 [3, 11, 49, 58]. Enzymatically active IRAK4 phosphorylates IRAK1, promoting its K48- and K63-linked ubiquitination, events that facilitate IRAK1 assembly with TRAF6 and IKK-γ, resulting in generation of modified IRAK1 species and disappearance of unmodified IRAK1 [3, 12, 15, 38, 49, 50, 53]. Therefore, it is not surprising that IRAK4 kinase deficiency resulted in a loss of LPS-mediated modifications of IRAK1. TNF-α, KC, and IL-12 p40 genes, as well as MAPK activation, are known to exhibit dual MyD88 and TRIF dependencies [42, 59, 60]. Engagement of TLR4 with LPS activates both pathways in mouse PMs and BMDMs, whereas Pam3Cys ligation of the TLR1–TLR2 complex triggers only the MyD88 pathway [3, 61, 62]. Thus, additional induction of the TRIF-dependent pathway by LPS, but not by Pam3Cys, could perhaps account for differences in how IRAK4 kinase deficiency affects TLR2- vs. TLR4-mediated activation of MAPK and cytokines. Consistent with this notion, we found that LPS-inducible expression of RANTES and IFN-β—genes that are highly dependent on the TRIF pathway for full induction [42, 63, 64]—was comparable in IRAK4 kinase-deficient macrophages compared with WT cells. These data indicate that IRAK4 kinase activity is important for TLR2/4-mediated activation of the MyD88 pathway but is dispensable for TRIF-dependent signaling.

To the best of our knowledge, our data are the first to indicate that deficiency in IRAK4 kinase activity impairs induction of TLR2 homotolerance and TLR2 → TLR4 heterotolerance at the level of JNK and p38 MAPK activation. In contrast, expression of kinase-inactive IRAK4 does not impact endotoxin tolerance, as evidenced by decreased LPS-inducible IκB-α degradation, ERK, JNK, and p38 MAPK activation; expression of cytokine genes; and induction of negative TLR regulators, IRAK-M and A20, despite the fact that IRAK4 deficiency results in a lower initial response to LPS. The preserved capacity of IRAK^KDKI macrophages to develop endotoxin tolerance, despite their reduced ability to produce MyD88-dependent cytokines, might be accounted for if IRAK4 kinase activity is dispensable for activation of the TRIF pathway, as demonstrated in this paper and reported elsewhere [15, 17, 39]. In this scenario, the loss of IRAK4 kinase activity ablates activation of TLR2-elicited, MyD88-dependent, IRAK4 kinase-driven pathways responsible for TLR2 homo- and cross-tolerance induction. In contrast, the preservation of IRAK4 kinase-independent, TRIF-driven pathways could explain the retained ability of IRAK4^KDKI macrophages to develop endotoxin tolerance. Notably, the TRIF-dependent pathway has been reported to be essential for LPS tolerance [65] and for LPS-mediated tolerance to ischemic injury [66].

In addition to differential use of the TRIF pathway, we cannot exclude differences in TLR4- versus TLR2-elicited, MyD88-dependent signaling that could be affected selectively by a deficiency in IRAK4 kinase activity. IRAK4 functions as a kinase, activating TAK1- and MEKK3-dependent signaling, and as an adapter, triggering MEKK3-dependent signaling that results in differential outcomes on NF-κB activation [14–16]. Indeed, TAK1-dependent signaling was reported to induce the classical pathway, resulting in activation of TAK1 and IKK-β, p-IκB-α and IκB-α protein degradation, and activation of NF-κB-dependent cytokine genes to the fullest extent [14]. On the other hand, kinase-inactive IRAK4 was found to activate MEKK3, p-IKK-γ, p-IκB-α but not IκB-α degradation, and induction of some but not all cytokine genes [12, 14, 15]. TLR2, TLR7, and TLR9 signaling exhibit different sensitivities to IRAK4 kinase activity, and TLR7/9 signaling is the most sensitive [15]. These findings somewhat echo our results, showing that IRAK4 kinase deficiency is associated with inhibition of TLR2-elicited MAPK activation and gene expression, impaired TLR2 homotolerance, and ablated TLR2 → TLR4 heterotolerance but has a milder impact on TLR4-mediated signaling and leads to retention of endotoxin tolerance. Although the exact molecular basis for this phenomenon remains unknown, one might speculate that TLR4 engagement induces a broader spectrum of the MyD88-dependent signaling cascades compared with a narrower repertoire induced through TLR2, leading to their differential sensitivity to IRAK4 functions as a kinase versus adapter. Intact activation of the TRIF pathway and induction of a broader spectrum of MyD88-dependent events could also account for preserved LPS inducibilty of IRAK-M and A20, negative regulators of TLR signaling linked to endotoxin tolerance [25, 28, 45–48].

Similar to previous reports [22, 24, 40], we showed herein that endotoxin tolerance ablates LPS-inducible expression of IFN-β in WT macrophages but does not affect RANTES expression. Likewise, IRAK4^KDKI macrophages showed the same trend, in line with our findings that IRAK4 kinase activity is dispensable for TLR4 tolerance. The exact molecular mechanisms responsible for differential sensitivity of IFN-β versus RANTES to endotoxin tolerance are unclear but could be related to a differential contribution of various transcription factors, coactivators, and corepressors. As micro-RNAs have been reported to regulate TLR signaling pathways (reviewed in refs. [67, 68]), we also cannot exclude differences in the induction of micro-RNAs in WT versus IRAK4^KDKI macrophages that could differentially impact the stability of IFN-β and RANTES mRNA. Future studies will be needed to discern the exact mechanisms responsible for differential sensitivity of IFN-β and RANTES to endotoxin tolerance.

Studies of human patients expressing IRAK4 mutations, predicted to yield kinase-deficient species, demonstrated high incidence of pyogenic infections but the ability to clear Gram-negative and viral infections [69, 70], similar to patients expressing mutations in MyD88 [71]. A pioneering study by Pennini et al [39] revealed a markedly higher susceptibility of IRAK4^KDKI mice compared with WT animals to infection with Streptococcus pneumonia, whereas treatment of mice with TLR3 agonists (activating the TRIF pathway) significantly improved survival in both groups. Paradoxically, despite accumulation of autoreactive B cells in the blood, IRAK4- and MyD88-deficient patients did not display autoreactive serum antibody or develop autoimmune diseases [72], suggesting that autoreactive B cells in their blood are not activated and do not secrete these antibodies when TLR signaling is attenuated. Further studies are required to delineate how IRAK4 kinase-dependent and independent pathways contribute to the pathogenesis of infectious and inflammatory diseases.

In summary, this represents one the first reports dissecting the role of IRAK4 kinase activity in TLR2 and TLR4 homotolerance and TLR2-mediated heterotolerance of TLR4. IRAK4 kinase activity is dispensable for endotoxin tolerance, as evidenced by suppressed p-ERK, p-JNK, and p-p38, IκB-α degradation, induction of proinflammatory cytokines, and up-regulation of negative regulators IRAK-M and A20. In contrast, IRAK4 kinase activity is critical for TLR2-elicited inhibition of Pam3Cys- and LPS-inducible p-JNK and p-p38 MAPKs. Studies are in progress to discern the molecular mechanisms by which IRAK4 kinase activity regulates TLR signaling, tolerance, and sensitivity to microbial infections, septic shock, and autoimmunity.

AUTHORSHIP

A.E.M. and S.N.V. conceived the idea, designed the experiments, oversaw the entire project, and prepared the manuscript with input from each coauthor. Y.X. and M.P. performed analyses of mRNA and protein expression in WT and IRAK4^KDKI macrophages and collected cell-free supernatants for ELISA.