Endotoxin Tolerance Inhibits Lyn and c-Src Phosphorylation and Association with Toll-Like Receptor 4 but Increases Expression and Activity of Protein Phosphatases

Abstract

Endotoxin tolerance protects the host by limiting excessive 'cytokine storm' during sepsis, but compromises the ability to counteract infections in septic shock survivors. It reprograms Toll-like receptor (TLR) 4 responses by attenuating the expression of proinflammatory cytokines without suppressing anti-inflammatory and antimicrobial mediators, but the mechanisms of reprogramming remain unclear. In this study, we demonstrate that the induction of endotoxin tolerance in human monocytes, THP-1 and MonoMac-6 cells inhibited lipopolysaccharide (LPS)-mediated phosphorylation of Lyn, c-Src and their recruitment to TLR4, but increased total protein phosphatase (PP) activity and the expression of protein tyrosine phosphatase (PTP) 1B, PP2A, PTP nonreceptor type (PTPN) 22 and mitogen-activated protein kinase phosphatase (MKP)-1. Chemical PP inhibitors, okadaic acid, dephostatin and cantharidic acid markedly decreased or completely abolished LPS tolerance, indicating the importance of phosphatases in endotoxin tolerization. Overexpression of PTPN22 decreased LPS-mediated nuclear factor (NF)-κB activation, p38 phosphorylation and CXCL8 gene expression, while PTPN22 ablation upregulated LPS-induced p65 NF-κB and p38 phosphorylation and the expression of TNF-a and pro-IL-1ß mRNA, indicating PTPN22 as an inhibitor of TLR4 signaling. Thus, LPS tolerance interferes with TLR4 signaling by inhibiting Lyn and c-Src phosphorylation and their recruitment to TLR4, while increasing the phosphatase activity and expression of PP2A, PTPN22, PTP1B and MKP1.

Introduction

During infection and inflammation, microbial components and endogenous ‘alarmins’ activate membrane-associated Toll-like receptors (TLRs) expressed on macrophages, dendritic cells and neutrophils to activate innate immunity and prime adaptive immune responses [1]. TLRs contain an ectodomain with multiple leucine-rich regions involved in ligand sensing and coreceptor interactions, a transmembrane domain and a cytoplasmic tail with a signaling Toll-IL-1R (TIR) domain [1, 2]. TLR2 cooperates with TLR1 or TLR6 to detect tri- and di-acylated lipoproteins expressed by Gram-positive bacteria, mycobacteria and mycoplasma [3, 4]. TLR4 is the main sensor of Gram-negative bacterial lipopolysaccharide (LPS) [5], but also detects the fusion protein of respiratory syncytial virus, mouse mammary tumor virus and human cytomegalovirus [1]. TLR5 recognizes extracellular flagellin, TLR3 and TLR7/8 sense viral double-stranded and single-stranded RNA, respectively, while TLR9 detects unmethylated CpG motifs present in bacteria, fungi and viruses [1, 6]. TLRs also sense endogenous ‘alarmins’, e.g. high-mobility group box-1 protein, released from cells during infection, inflammation and stresses [reviewed in [1, 7, 8]].

Ligand recognition initiates TLR dimerization that brings together TIR domains, leading to the assembly of docking platforms to enable recruitment of downstream adapters and kinases [9, 10]. The majority of TLRs recruit myeloid differentiation primary response protein (MyD) 88 and IL-1R-associated kinase (IRAK)-4, IRAK-1 and IRAK-2 to signal the activation of transcription factors, i.e. nuclear factor (NF)-κB, leading to the expression of inflammatory cytokines [10]. TLR3 recruits TIR domain-containing adapter-inducing interferon (IFN)-β (TRIF) to induce proinflammatory cytokines via receptor-interacting protein-1, and uses TRIF-TNFR-associated factor (TRAF) family member-associated NF-κB activator-binding kinase-1 (TBK-1) to activate IFN-regulatory factors (IRF)-3 and IRF7, leading to the induction of type I IFNs [1]. TLR4 induces proinflammatory cytokines from the cell surface via the MyD88-IRAK pathway and signals from the endosomal compartment via the TRIF-TBK1-IRF3 module to activate type I IFNs [11].

Endotoxin tolerance is a state of reprogramming of TLR4 responses to LPS following prior exposure to LPS [12, 13, 14, 15]. It is hypothesized to protect the host against excessive inflammatory responses, e.g. in septic patients surviving an initial cytokine storm [13, 16, 17, 18], but renders the host immunocompromised and susceptible to secondary infections [13, 16]. We previously identified impaired LPS-driven tyrosine phosphorylation of TLR4 and MyD88 adapter-like (Mal) as hallmarks of endotoxin tolerance [19, 20] and suggested protein tyrosine kinases (PTK) of the Src family as affected molecules [19]. This paper shows that endotoxin tolerance inhibits LPS-inducible phosphorylation of Lyn and c-Src on tyrosines in catalytic loops associated with kinase activation, impairs Lyn and c-Src recruitment to TLR4, increases total protein phosphatase (PP) activity and expression of PP2A, protein tyrosine phosphatase (PTP) 1B, PTP nonreceptor type (PTPN) 22 and mitogen-activated protein kinase (MAPK) phosphatase (MKP)-1. Phosphatase chemical inhibitors, okadaic acid (OA), dephostatin (DP) or cantharidic acid (CA) blunted or completely abolished endotoxin tolerization, indicating the critical importance of phosphatases in LPS tolerance. Overexpression of PTPN22 attenuated LPS-driven NF-κB activation, p38 MAPK phosphorylation and CXCL8 gene expression, while PTPN22 ablation increased LPS-mediated p65 NF-κB and p38 phosphorylation and the induction of TNF-α and pro-IL-1β mRNA, suggesting that increased PTPN22 in LPS-tolerized macrophages can act as an inhibitor of TLR4 signaling.

Materials and Methods

Reagents and Cell Culture

Highly purified, protein-free Escherichia coli K235 LPS was prepared as described previously [21]. The following Abs were used: anti (α)-phospho (p) and α-total p38, α-PTPN22, α-PP2A subunit A, α-IκB-α, α-total and α-p (pY416)-c-Src, α-total Lyn (Cell Signaling Technology, Danvers, Mass., USA), α-TLR4, α-β-actin, α-PTP1B, α-MKP-1, α-p-p65 and α-tubulin (Santa Cruz Biotechnology, Santa Cruz, Calif., USA), and α-p (Y396)-Lyn (Abcam, Cambridge, Mass., USA). OA, DP and CA were from Sigma-Aldridge (St. Louis, Mo., USA). Human monocytes were kindly provided by Dr. Larry Wahl (NIDCR/NIH) as deidentified samples prepared by counter flow elutriation of blood from healthy human volunteers [22]. The myelomonocytic cell lines THP-1 were from the ATCC (Manassas, Va., USA), and MonoMac-6 cells [23] were kindly provided by Dr. Jorge Cervantes (University of Connecticut Health Center, Farmington, Conn., USA). THP-1, MonoMac-6 cells and monocytes were cultured in RPMI 1640 containing 2 mML-glutamine, 100 μg/ml penicillin, 100 μg/ml streptomycin (Life Technologies, Grand Island, N.Y., USA), 5 × 10^-5M β-mercaptoethanol and 5% HyClone FBS (Thermo Scientific, Waltham, Mass., USA). For all functional assays, THP-1 cells were differentiated for 72 h with 20 ng/ml PMA to attain macrophage characteristics, e.g. plastic adherence, CD14 expression and phagocytosis [24]. 293/YFP-TLR4 cells were described [25] and cultured in DMEM supplemented with 2 mML-glutamine, 100 μg/ml penicillin, 100 μg/ml streptomycin and 10% FBS. Studies with human monocytes were approved by the Institutional Review Board of the University of Connecticut Health Center and University of Maryland School of Medicine. pEFBOS-MD2, pcDNA3-CD14, pE-selectin leukocyte adhesion molecule (ELAM)-luciferase, p-thymidine kinase (TK)-Renilla luciferase reporters (Promega) and pCMV2B vectors encoding Flag-PTPN22 were as described [19, 20, 26, 27].

Transfection and Lentivirus Transduction

293/YFP-TLR4 cells were plated in 24-well plates (reporter assays) or 6-well plates (RNA and protein assays) and transfected with pcDNA3-CD14, pEFBOS-MD2 without (RNA and protein assays) or with (reporter assays) pELAM-luciferase and pTK-Renilla luciferase, using lipofectamine 2000 transfection reagent (Life Technologies) as recommended by the manufacturer and described previously [19, 20, 26]. After recovery for 48 h, cells were washed and treated with medium or LPS, as indicated. For lentivirus transduction, THP-1 cells were incubated with ready-to-use lentivirus particles encoding scrambled or PTPN22 shRNA (Santa Cruz) according to the manufacturer's protocol, and stably transduced cells were selected for 2 weeks with 5 μg/ml puromycin.

Measurement of Total PP Activity

Cells in 6-well plates (2 × 10⁶ cells/well) were treated with medium or 100 ng/ml LPS, washed in PBS, resuspended in an ice-cold lysis buffer containing 50 mM Tris-HCl (pH 7.0), 0.1 mM EDTA, 25 μg/ml aprotinin, 25 μg/ml leupeptin and 25 μg/ml PMSF, incubated for 20 min and centrifuged (15,000 g, 15 min, 4°C). The protein concentration was measured using a DC Protein Assay kit (Bio-Rad). Total phosphatase activity was determined as described previously [28]. In brief, cell extracts were incubated for 60 min at 37°C in reaction mixtures containing 10 mM p-nitrophenyl-phosphate, 116 mM NaCl, 5.4 mM KCl, 5.5 mMD-glucose and 50 mM HEPES (pH 7.0) with gentle shaking. The reactions were terminated by the addition of 200 μl of 1 M NaOH followed by measuring absorbance at 405 nm. The results were normalized to 10 μg of protein.

RNA Isolation and Quantitative Real-Time PCR (RT-qPCR) Analysis

RNA was isolated with Trizol (Life Technologies), treated with DNase and repurified. cDNA was prepared from 1 μg of total RNA using a reverse transcription system (Promega) and examined by RT-qPCR, using 5 µl of cDNA, 0.3 μM of gene-specific primers and SYBR Green Supermix (Bio-Rad) on a MyIQ RT-PCR machine (Bio-Rad). The following primers were used: human hypoxanthine phosphoribosyl transferase (HPRT), forward: 5′-ACCAGTCAACAGGGGACATAAAAG-3′, reverse: 5′-GTCTGCATTGTTTTGCCAGTGTC-3′; TNF-α, forward: 5′-CCCAGGCAGTCAGATCATCTTCC-3′, reverse: 5′-GCTTGAGGGTTTGCTACAACATG-3′; CXCL-8, forward: 5′-CACCGGAAGGAACCATCTCACT-3′, reverse: 5′-TGCACCTTCACACAGAGCTGC-3′; pro-IL-1β, forward: 5′-AAATGATGGCTTATTACAGTGGCA-3′, reverse: 5′-CTTCATCTGTTTAGGGCCATCAG-3′; MKP-1, forward: 5′-CAGTACCCCACTCTACGATCA-3′, reverse: 5′-ACCCTCAAAATGGTTG GGACA-3′; PTP1b, forward: 5′-GCTTCTCCTACCTGGCTGTG-3′, reverse: 5′-TTGTGTGGCTCCAGGATTC-3′; PP2A, forward: 5′-TGCCAATGGTCTCACACTGG-3′, reverse: 5′-GCCTGGTTCCC ACAACGATA-3′; PTPN22, forward: 5′-CCACTTCCTGTACGGACACC-3′, reverse: 5′-TGTTCCACCCCATTCCAGTG-3′. Data were analyzed as previously described [19, 20, 26, 29].

Preparation of Cell Extracts, Immunoprecipitation and Western Blot Analyses

Cells were lysed for 30 min in a 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 Complete™ protease inhibitor cocktail (Roche Applied Science, Indianapolis, Ind., USA). Supernatants were collected and the protein concentration was determined using the DC Protein Assay kit (Bio-Rad). Cell lysates were resuspended in Laemmli sample buffer (Bio-Rad) and boiled for 10 min. For immunoprecipitation, lysates (500 μg) were precleared for 2 h with 20 μl of protein G agarose (50% slurry) and incubated for 20 h at 4°C with Abs (1 μg). Immune complexes were pulled down for 4 h with 40 μl of protein G-agarose, beads were washed 5 times in lysis buffer, boiled in Laemmli sample buffer and centrifuged, and the supernatants were analyzed. Proteins were separated on 4–20% polyacrylamide gels (Life Technologies) and electrotransferred to Immobilon-P membranes (Bio-Rad). Membranes were blocked in TTBS/5% milk, probed with Abs, washed, incubated with secondary IgG-horseradish peroxidase, and the bands were visualized by enhanced chemiluminescence.

NF-κB Reporter Assays

Cells were transfected with plasmids, using Lipofectamine 2000 transfection reagent according to the manufacturer's recommendations, recovered for 24 h and treated for 5 h with medium or 100 ng/ml LPS. Cells were lysed in a passive lysis buffer (Promega), and firefly luciferase versus Renilla luciferase activities were measured using the dual luciferase reporter assay system (Promega) on a Berthold LB 9507 luminometer (Berthold Technologies).

Statistical Analysis

Data were analyzed with the GraphPad Prism 5 program for Windows (GraphPad Software, San Diego, Calif., USA), using the Student t test with the confidence interval set at 95% for pairwise comparisons, and a one-way ANOVA with repeated measures, followed by post hoc comparisons with Tukey's multiple paired comparison test to compare multiple groups. The results are expressed as the mean ± SD values.

Results

Induction of Endotoxin Tolerance in Human Macrophages Attenuates LPS-Inducible Recruitment to TLR4 and Modulates Phosphorylation of Lyn and c-Src

We reported reduced LPS-mediated tyrosine phosphorylation of TLR4 and Mal in LPS-tolerant human monocytes [19, 20] and showed that endotoxin tolerization of 293/TLR4/MD2 cells inhibits recruitment of transfected Flag-Lyn to TLR4 in response to LPS [19]. In the present study, we examined the impact of endotoxin tolerance on LPS-induced phosphorylation of endogenous Lyn and c-Src and recruitment of these kinases to TLR4 in human THP-1 cells differentiated with PMA to attain the macrophage phenotype (plastic adherence, CD14 expression, phagocytosis [24]). As shown in figure 1, LPS stimulation of medium-pretreated THP-1 cells induced the recruitment of c-Src (fig. 1a) and Lyn (fig. 1b) to TLR4 and led to their phosphorylation on Y416 (c-Src) and Y396 (Lyn), correlating with cell activation as evidenced by phosphorylation of p38 MAPK. Prior exposure to LPS induced endotoxin tolerance, as evidenced by ablated LPS-induced p38 phosphorylation (fig. 1). Endotoxin-tolerized THP-1 macrophages exhibited attenuated recruitment of c-Src and Lyn to TLR4 (fig. 1a, b), top panels) and reduced the phosphorylation of these PTKs in response to LPS challenge. Comparable expressions of total c-Src, Lyn and TLR4 were detected in control (medium-pretreated) and LPS-tolerized monocytes and THP-1 cells, indicating that differences in the extent of PTK phosphorylation and recruitment to TLR4 were not due to variations in their expression levels. Similar levels of β-actin (fig. 1a) or tubulin (fig. 1b) were seen in the samples analyzed, showing that differences in c-Src and Lyn phosphorylation could not be accounted for by unequal loading. These results show that endotoxin tolerance attenuates LPS-induced phosphorylation of c-Src and Lyn and their association with TLR4.

Fig. 1.

Induction of endotoxin tolerance inhibits LPS-induced phosphorylation of c-Src and Lyn and their recruitment to TLR4. THP1 cells were treated for 20 h with medium or 100 ng/ml LPS, washed three times and challenged with medium (-) or 100 ng/ml LPS for the indicated times. Whole cell lysates were immunoprecipitated with Abs against c-Src (a) or Lyn (b), followed by Western blot analyses with α-TLR4 Ab or immunoblotted with the indicated Abs. The results of a representative experiment (n = 3) are depicted.

LPS-Tolerized Human Monocytes, THP-1 and MonoMac-6 Cell Lines Exhibit Increased Expression of PPs MKP-1, PTP1b, PP2A and PTPN22

It is plausible that impaired phosphorylation of TLR4, Mal and Src kinases (Lyn and c-Src) in endotoxin-tolerant human macrophages could be due to increased levels of PPs. To begin addressing this question, we determined the impact of endotoxin tolerance on mRNA and protein expression of several phosphatases in human monocytes and macrophage-like cell lines. We chose to analyze phosphatases implicated in dephosphorylation of p38 MAPK (MKP-1) [30, 31], disruption of TLR4-MyD88 interactions (PP2A) [32], negative regulation of MyD88- and TRIF-dependent cytokines (PTP1B) [33], regulation of macrophage M1-M2 polarization [27] and activation of TRAF-3 and type I IFNs (PTPN22) [34]. LPS pretreatment of human monocytes, THP-1 and MonoMac-6 cells led to 88–95% decreases in the induction of TNF-α mRNA upon subsequent LPS challenge compared to responses seen in medium-pretreated, LPS-restimulated cells, documenting endotoxin tolerance (fig. 2a, b, c, left columns). Despite decreased LPS-induced TNF-α mRNA levels, endotoxin-tolerized human monocytes, THP-1 and MonoMac-6 cells exhibited 2.1- to 5.0-, 1.5- to 4.0-, 3.3- to 30.0- and 2.5- to 6.1-fold increases in the expression of PP2A, PTPN22, PTP1B and MKP-1 mRNA, respectively, compared to the levels of these phosphatases in control cells, which remained persistent and increased upon LPS challenge (fig. 2a, b, c, 2nd and 3rd panels from the left, and data not shown).

Fig. 2.

Endotoxin-tolerized human monocytes and macrophages exhibit increased levels of MKP-1, PTP1B, PP2A and PTPN22 mRNA. Human monocytes (a), THP-1 (b) and MonoMac-6 cells (c) were exposed for 20 h to medium or 100 ng/ml LPS. After washing, cells were treated for 3 h with medium or 100 ng/ml LPS; RNA was isolated, reverse-transcribed and subjected to real-time qPCR analyses with the corresponding gene-specific primers. Data are presented as the fold increase compared to medium-exposed cells treated with medium. The results (mean ± SD) of a representative experiment (n = 4) are shown. * p < 0.05.

To extend our data, we next examined the impact of endotoxin tolerization on the total protein levels of the analyzed PPs. Induction of endotoxin tolerance (evidenced by impaired LPS-driven p38 phosphorylation) led to increased expression of PP2A, PTPN22, PTP1B and MKP-1 proteins in human monocytes (online suppl. fig. S1; see www.karger.com/doi/10.1159/000440838 for all online suppl. material, and data not shown), THP-1 and MonoMac-6 cells (fig. 3, online suppl. fig. S2) compared to their levels in medium-pretreated cells. Increased expression of these PPs in endotoxin-tolerized cells was not due to higher protein loading of these samples, as we observed equal or even lower intensities of the tubulin and total p38 bands in the samples of LPS-tolerant cells compared to nontolerized controls (fig. 3, online suppl. fig. S2). These data indicate that endotoxin tolerization of human monocytes and macrophage-like cell lines increases the expression of PTP1B, MKP-1, PP2A and PTPN22 phosphatases.

Fig. 3.

Endotoxin tolerization of THP1 and MonoMac-6 cells upregulates protein expression of PP2A, PTPN22, PTP1B and MKP-1. THP1 (a) and MonoMac-6 (b) cells were pretreated for 20 h with medium or 100 ng/ml LPS, washed three times and restimulated with medium (0 time point) or 100 ng/ml LPS for the indicated times. Whole cell lysates were analyzed by Western blotting with Abs against PP2A, PTPN22, PTP1B, MKP-1, p-p65, p-p38, total p38 and tubulin. The results of a representative experiment (n = 4) are shown.

Endotoxin Tolerization Increases Total PP Activity and Inhibition of PPs Attenuates the Induction of Endotoxin Tolerance

Next, we examined whether the increased expression of PPs observed in endotoxin-tolerant macrophages results in the upregulation of total phosphatase activity. THP-1 cells exhibited 2- and 3-fold increases in phosphatase activity as early as 1 and 3 h post-LPS treatment, which reached a 4-fold increase by 20 h (endotoxin tolerance induction) and was further increased to a 6-fold increase in endotoxin-tolerized cells treated with LPS for 3 h (fig. 4). LPS stimulation of endotoxin-tolerized THP-1 cells for 3 h further increased phosphatase activity, reaching an approximately 7-fold increase compared to medium-treated cells (fig. 4).

Fig. 4.

Induction of endotoxin tolerance upregulates total PP activity. THP1 cells pretreated for 20 h with medium or 100 ng/ml LPS were washed and challenged with medium or 100 ng/ml LPS for the indicated times. Whole cell lysates were prepared and PP activity was analyzed. Data (mean ± SD) of a representative experiment (n = 3) are presented. * p < 0.05.

To mechanistically address the role of phosphatase activity in the induction of endotoxin tolerance, we used chemical inhibitors of PPs. These included OA, a general chemical inhibitor of phosphatase activity [28], DP, an inhibitor of several PTPs, including CD45, SHP-1 and PTP1B [35, 36], and CA, a chemical inhibitor of PP2A and PP1 [37]. OA at concentrations 10 and 20 nM exerted a dose-dependent inhibition of LPS-induced total phosphatase activity in THP-1 cells by 65–80%, whereas at 5 nM OA did not influence this response (fig. 5a). In the absence of OA, LPS pretreatment reduced the capacity of THP-1 cells to express TNF-α mRNA in response to subsequent LPS challenge by 95% (fig. 5b). The addition of 20 nM of OA during pretreatment reduced the extent of endotoxin tolerization by approximately 2-fold, as evidenced by a decrease in the suppression of LPS-mediated TNF-α gene expression from 95 to 46% (fig. 5b). To substantiate these results, we studied the impact of other PP chemical inhibitors, CA (53 μM) and DP (9 μM), on the induction of endotoxin tolerance at the doses reported to efficiently decrease PP activity [38, 39, 40]. Prior exposure of THP-1 cells with LPS in the presence of CA (fig. 5c) or DP (fig. 5d) led to almost complete restoration of their LPS inducibility of TNF-α gene expression comparable to that observed in medium-pretreated cells. Neither of the chemical inhibitors affected viability of THP-1 cells (93–97% viability as determined by trypan blue exclusion). Taken collectively, these data suggest that LPS-driven phosphatase activity is an important regulatory mechanism controlling the extent of endotoxin tolerization and TLR4 signaling.

Fig. 5.

OA, DP and CA reduce or reverse the induction of endotoxin tolerance in THP1 cells. a THP-1 cells were treated for 24 h with medium or 100 ng/ml LPS in the absence (0) or in the presence of the indicated concentrations of OA, and the phosphate concentration in the tissue culture supernatant was determined. The data (mean ± SD) of a representative experiment out of three total experiments are depicted. THP1 cells were exposed to medium or 10 ng/ml LPS with or without 20 nM OA (b), 9 μM DP (c) or 53 μM CA (d). Thereafter, the cells were washed, treated for 3 h with medium or 100 ng/ml LPS, RNA was isolated, reverse-transcribed and subjected to real-time qPCR analyses of HPRT and TNF-α mRNA levels. The results (mean ± SD) of a representative experiment (n = 3) are shown and the percentages of inhibition are given above the dashed lines. * p < 0.05

PTPN22 Acts as an Inhibitor of LPS-Driven p38 Phosphorylation, NF-κB Activation and Cytokine Gene Expression

While PP2A and MKP-1 ablation experiments have established these phosphatases as important control checkpoints of TLR4 signaling [30, 32, 41, 42], the role of PTPN22 in mouse myeloid cells is less clear, and its functions in human monocytes and macrophages remain to be deciphered. In mice, PTPN22 has been shown to suppress the expression of proinflammatory cytokines such as IL-12, IL-6 and IL-1β, but to enhance the production of type I IFN by myeloid cells in response to LPS [27, 34]. To further address the role of PTPN22 in the modulation of TLR4 signaling in human cells, we first mimicked enhanced PTPN22 levels by transfection of 293/TLR4 cells with pCMV2b plasmid encoding Flag-PTPN22, along with vectors encoding CD14 and MD2 that enable optimal LPS sensitivity, and analyzed LPS-driven p65 NF-κB and p38 MAPK phosphorylation, activation of cotransfected NF-κB luciferase reporter and induction of CXCL8 mRNA. Figure 6 shows that overexpression of WT PTPN22 led to a marked inhibition of LPS-driven phosphorylation of p65 NF-κB and p38 MAPK (fig. 6a) and dose-dependent suppression of LPS-mediated induction of NF-κB-dependent luciferase reporter (pELAM-luciferase) activation by 30–50% (fig. 6b) compared to robust LPS responses observed in cells transfected with empty vector. PTPN22 overexpression did not induce p38 MAPK phosphorylation or NF-κB activation without LPS stimulation (fig. 6a, b). To extend these results, we next determined the impact of PTPN22 overexpression on LPS-mediated induction of CXCL8 mRNA, which is controlled in an NF-κB-dependent manner [43]. Transfection of PTPN22 caused the dose-dependent inhibition of LPS-mediated induction of CXCL-8 gene expression by 18–38% (fig. 6c).

Fig. 6.

PTPN22 inhibits LPS-induced p38 MAPK phosphorylation, NF-κB activation and expression of CXCL8, TNF-α and pro-IL-1β cytokine genes. a, c 293/YFP-TLR4 cells plated in 6-well plates (3 × 10⁶ cells per well) were cotransfected with pEFBOS-MD2 and pcDNA3-CD14 (0.5 μg/well each), pCMV2 or pCMV2-Flag-PTPN22 (0.2 or 0.8 μg/well), and total plasmid input was adjusted to 2 μg/well with pcDNA3. a Cells were treated for the indicated times with medium or 100 ng/ml LPS, and cell lysates were subjected to Western blot analyses with the indicated Abs. b 293/YFP-TLR4 cells were plated in 24-well plates and cotransfected with pEFBOS-MD2, pcDNA3-CD14 (50 ng per well each), pELAM-luciferase (300 ng per well), pTK-Renilla luciferase (20 ng per well), pCMV2 or pCMV2-Flag-PTPN22. After recovery for 24 h, cells were treated for 5 h with medium or 100 ng/ml LPS, and firefly versus Renilla luciferase activities were determined in the cell lysates. The results (mean ± SD) of a representative experiment (n = 5) are presented as the NF-κB fold increase compared to samples transfected with pCMV and treated with medium (taken as 1). c Cells were treated for 3 h with medium or 100 ng/ml LPS, RNA was isolated, reverse transcribed and analyzed by real-time qPCR with the corresponding primers for human HPRT and CXCL-8. Data (mean ± SD) of a representative experiment (n = 3) are shown. * p < 0.05. d, e THP-1 cells were transduced with lentiviral particles encoding scrambled or PTPN22 shRNA, selected with puromycin for 2 weeks, differentiated for 72 h with 20 ng/ml PMA, washed and treated with medium or LPS for the indicated times (d, left panel, e) or for 3 h (d, middle and right panels). RNA was isolated, reverse transcribed and analyzed by real-time qPCR. Cell lysates were subjected to Western blot analyses to examine the expression of the indicated proteins (e, left and middle panels), and the results were quantified and expressed as arbitrary units (e, right panel). The results (mean ± SD) of a representative experiment (n = 3) are shown. * p < 0.05.

Secondly, we studied whether ablation of PTPN22 affects LPS-mediated p38 and NF-κB activation and induction of TNF-α and pro-IL-1β gene expression in PMA-differentiated THP-1 human macrophages transduced with scrambled or PTPN22-specific shRNAs. Lentiviral transduction of THP-1 cells with PTPN22 shRNA reduced the levels of endogenous PTPN22 mRNA by 68–86% and efficiently downregulated PTPN22 protein expression compared to cells transduced with scrambled shRNA, showing efficient silencing (fig. 6d). PTPN22 knockdown in THP-1 cells markedly upregulated LPS-induced phosphorylation of p38 MAP kinase and p65 NF-κB, and increased LPS-induced levels of TNF-α and pro-IL-1β mRNA by approximately 3-fold versus the response exhibited by control cells transduced with scrambled shRNA (fig. 6e). Collectively, these data suggest that increased PTPN22 expression during the induction of endotoxin tolerance inhibits LPS-mediated phosphorylation of p38 and activation of NF-κB, and suppresses proinflammatory cytokines and chemokines, e.g. CXCL8 and TNF-α.

Discussion

Our previous studies identified impaired LPS-induced tyrosine phosphorylation of TLR4 and Mal as molecular hallmarks of endotoxin tolerance and found that tyrosine mutants of TLR4 and Mal are signaling deficient and act as dominant negative inhibitors of TLR4 signaling [19, 20]. Using overexpression approaches in 293/TLR4/MD2 cells, we also showed inhibited LPS-driven recruitment of transfected Flag-Lyn to TLR4 in endotoxin-tolerized 293/TLR4/MD2 cells [19]. This paper provides several lines of evidence documenting the importance of phosphorylation and dephosphorylation events in the regulation of TLR4 signaling and induction of endotoxin tolerance in primary human monocytes, THP-1 and MonoMac-6 macrophage-like cell lines. It demonstrates that endotoxin tolerization of human macrophages leads to deficient LPS-induced phosphorylation of endogenous c-Src and Lyn on Y416 and Y396, respectively, correlating with deficient phosphorylation of p38 and activation of TNF-α gene expression. Since phosphorylation of c-Src and Lyn on Y416 (c-Src) and Y396 (Lyn) in their kinase loops induces kinase activities [44, 45], and because LPS-tolerized THP-1 cells showed inhibited recruitment of c-Src and Lyn to TLR4, it is plausible that deficient abilities of these kinases to associate and phosphorylate TLR4 are responsible for the impaired TLR4 tyrosine phosphorylation and signaling.

Secondly, we showed that decreased activation and TLR4 assembly of c-Src and Lyn was not due to their diminished expression but correlated with increased mRNA and protein levels of phosphatases MKP-1, PTP1B, PP2A and PTPN22 in endotoxin-tolerized human monocytes, THP-1 and MonoMac-6 cells. To the best of our knowledge, this is the first demonstration of increased expression of PTP1B, PP2A and PTPN22 phosphatases in LPS-tolerant human monocytes and macrophages. In line with increased expression of PP2A, PTP1B, PTPN22 and MKP-1 phosphatases following the induction of endotoxin tolerance, we found significantly elevated and sustained total PP activity in endotoxin-tolerized THP-1 cells. Our data on increased MKP-1 gene expression and total phosphatase activity in human monocytes and macrophages confirm and extend similar findings obtained in mouse macrophages and THP-1 cells rendered tolerant to LPS [28, 41]. To further delineate the role of PPs in endotoxin tolerance, we used several chemical inhibitors that either generally target PPs (OA) or preferentially affect PP2A/PP1 (CA) or PTP1B/SHP-1 (DP). This paper demonstrates that the addition of chemical inhibitors during LPS pretreatment of THP-1 cells significantly blunted (OA) or completely reversed (DP, CA) endotoxin tolerization of THP-1 cells, as judged by LPS-inducible TNF-α mRNA. Together, these results indicate a critical role of PPs in the regulation of TLR signaling and induction of endotoxin tolerance.

Our findings that endotoxin tolerization upregulates several phosphatases, including PTP1B, PP2A, MKP-1 and PTPN22, may suggest a multilayered, coordinated control of TLR signaling exerted by these enzymes. PTP1B has been reported to inhibit both MyD88-dependent and TRIF-dependent signaling outcomes, e.g. activation of MAP kinases, NF-κB, IRF-3 and production of TNF-α, IL-6 and IFN-β in macrophages stimulated with TLR3, TLR4 and TLR9 agonists [33]. However, the molecular mechanisms by which PTP1B negatively regulates TLR signaling and how its deficiency affects capacities of monocytes and macrophages to become endotoxin tolerant are unknown. PTP1B can exert its effects by directly associating with TLRs, or it could target a signaling mediator involved in both MyD88- and TRIF-dependent pathways, such as TRAF-6 [1, 10]. Studies are in progress to address this question. In contrast to PTP1B affecting both the MyD88 and TRIF pathways, PP2A catalytic subunit α has been found to primarily affect MyD88-dependent signaling in LPS-tolerant mouse macrophages [32]. PP2A acts via disruption of the signal-promoting TLR4-MyD88 complex, promoting nuclear translocation of PP2A-MyD88 and inducing immunosuppressive patterns of chromatin modifications, such as H3S10 dephosphorylation [32]. Our results indicate significant upregulation of PP2A expression both at the mRNA and protein level, and LPS-tolerant RAW cells were reported to have an increased association of PP2A catalytic subunit α with MyD88 without significant changes in PP2A expression [32]. These findings suggest that different control mechanisms operate in mouse and human macrophages, which is in line with previous reports documenting the differences between these macrophages [46, 47]. Our previous work indicated that endotoxin tolerance inhibits TLR4-MyD88 association and ablates the activation of IRAK4 and IRAK1 [22, 26, 48, 49], and recent findings by Xie et al. [32] demonstrated that PP2A interferes with TLR4-MyD88 assembly and associates with MyD88. Given these findings, it is tempting to speculate that PP2A directly dephosphorylates components of the upstream signalosome, such as TLR4 or Mal, or deactivates kinases associated with the TLR4-Mal-MyD88 receptor complex. MKP-1 plays an important role in host defense, sepsis and endotoxin tolerance as MKP-1-deficient mice exhibit an exaggerated production of proinflammatory cytokines during infection and endotoxemia, an increased sensitivity to endotoxin shock [30, 42, 50] and fail to become endotoxin tolerized. MKP-1 is known to regulate the activation of several MAP kinases, including JNK and p38, via dephosphorylation, disabling their signaling potential [50, 51], suggesting JNK and p38 as primary targets of MKP-1.

PTPN22 has been recognized as a critical negative regulator of TCR signaling, and recent work has also revealed that it controls TLR signaling [27, 34]. However, its role in endotoxin tolerance has not been established, and the molecular mechanisms by which PTPN22 regulates TLR signaling are still poorly defined. This paper shows for the first time that increased PTPN22 mRNA and protein expression represents a novel hallmark of endotoxin tolerance in human monocytes and macrophage-like THP-1 and MonoMac-6 cells. Furthermore, we found significant inhibition of LPS-driven p38 MAPK phosphorylation, NF-κB activation and CXCL8 gene expression following transient transfection of PTPN22 in 293/TLR4/MD2 cells, consistent with our published results in PTPN22^−/− mice, demonstrating that PTPN22 inhibited the expression of inflammatory cytokines by LPS/IFN-γ-stimulated macrophages [27]. Conversely, PTPN22 knockdown in THP-1 cells led to a marked increase in LPS-mediated phosphorylation of p65 NF-κB, p38 MAPK and induction of TNF-α and pro-IL-1β mRNA, further demonstrating the inhibitory role for PTPN22 in TLR4 signaling. Mechanisms by which PTPN22 regulates TLR signaling are poorly defined, and controversial data have been reported in mice and mouse macrophages. While Wang et al. [34] reported that PTPN22 primarily affects LPS- and poly (I:C)-mediated induction of TRIF-dependent ubiquitination of TRAF-3, phosphorylation of IRF-3 and expression of type I IFNs, we found that PTPN22 regulates macrophage polarization and the expression of proinflammatory mediators [27]. Our ongoing studies seek to establish molecular targets of PTPN22 in human macrophages and to determine whether PTPN22 acts in concert with other PTPs to mediate TLR reprogramming upon LPS tolerization.

In summary, this study provides new results on the role of phosphorylation and recruitment to TLR4 of endogenous PTKs, Lyn and c-Src, and establishes the increased expression of several PPs, e.g. PTP1B, PP2A, MKP-1 and PTPN22, and enhanced total phosphatase activity as critical regulatory circuits controlling TLR4 signaling and endotoxin tolerance. Based on our data, it is tempting to speculate that tyrosine and Ser/Thr phosphatases could act in concert and coordinately dephosphorylate TLR4 and Mal within tyrosine and serine/threonine residues, leading to the reprogramming of TLR4 signaling in endotoxin-tolerized human monocytes and macrophages. Further studies are under way to explore this hypothesis.

Supplementary Material

Supplementary data

Supplementary data

Supplementary data