Abstract
Neutrophils influence innate and adaptative immunity by generating numerous mediators whose regulation largely depends on the IκB kinase (IKK)/IκB/NF-κB signaling cascade. A singular feature of neutrophils is that they express several components of this pathway (namely, NF-κB/Rel proteins and IκB-α) in both the nucleus and cytoplasm. We recently reported that the IKK complex of neutrophils is similarly expressed and activated in both cellular compartments. However, the upstream IKK kinase has not yet been identified. In this study, we report that neutrophils express the mitogen-activated protein 3 kinase, TGF-β–activated kinase 1 (TAK1), as well as its associated partners, TAK1-binding protein (TAB) 1, TAB2, and TAB4, in both the cytoplasm and nucleus. Following cell stimulation by TNF-α or LPS, TAK1 becomes rapidly and transiently activated. Blocking TAK1 kinase activity with a highly selective inhibitor (5z-7-oxozeaenol) attenuated the phosphorylation of nuclear and cytoplasmic IKKα/β, IκB-α, and RelA, and also impaired IκB-α degradation and NF-κB DNA binding in activated neutrophils. Moreover, TAK1 was found to be involved in the activation of p38 MAPK and ERK, which also influence cytokine generation in neutrophils. As a result, inflammatory cytokine expression and release were profoundly impaired following TAK1 inhibition. Similarly, the delayed apoptosis observed in response to LPS or TNF-α was reversed by TAK1 inhibition. By contrast, IKKγ phosphorylation and STAT1 activation were unaffected by TAK1 inhibition. Our data establish the central role of TAK1 in controlling nuclear and cytoplasmic signaling cascades in primary neutrophils, making it a promising target for therapeutic intervention in view of the foremost role of neutrophils in several chronic inflammatory conditions.
The NF-κB/Rel proteins are ubiquitous transcription factors that play a crucial role in immunity and inflammation (reviewed in Ref. 1). In resting cells, NF-κB complexes are usually sequestered in the cytoplasm through their interaction with inhibitory IκB proteins (2–5). Most of the signals leading to NF-κB activation converge on the activation of an IκB kinase (IKK) complex, which comprises at least three subunits (6). The first two (IKKα and IKKβ) are related catalytic subunits, whereas IKKγ is a regulatory subunit. Activation of the IKK complex involves the phosphorylation of all three subunits, leading to the degradation of IκB-α and nuclear translocation of NF-κB dimers, an essential event for the transcription of several early-response genes encoding cytokines and chemokines.
The molecular mechanism(s) by which the IKK complex becomes activated remain to be fully elucidated. In this regard, at least two models have been proposed. The first suggests that IKK recruitment to surface receptor complexes results in its autophosphorylation and subsequent activation (1). The other model involves phosphorylation and activation of the IKKs by a mitogen-activated protein 3 kinase (MAP3K), such as MEK kinase (MEKK) 1 (7), MEKK3 (8, 9), TANK-binding kinase 1 (10), and/or TGF-β–activated kinase 1 (TAK1) (11–13). Among these various MAP3Ks, however, TAK1 increasingly emerges as the most central IKK kinase (14). Indeed, knockout studies have shown that TAK1 gene deletion leads to severely impaired IKK and NF-κB activation by several stimuli (15, 16). Investigation of TAK1 activation has revealed that the kinase functions in association with its binding proteins (e.g., TAK1-binding protein [TAB] 1, TAB2) (17–19), which can be recruited to the IKK complex by TNFR-associated factor (TRAF)-2/6, receptor interacting protein (RIP) 1, and/or IKKγ (13, 20–25). This allows the TAK1-mediated phosphorylation of IKKβ, resulting in a catalytically active IKK complex (21).
Neutrophils are best known as professional phagocytes, but can also be an important source of NF-κB–dependent cytokines and chemokines (26, 27). A singular feature of neutrophils is that they constitutively express large amounts of NF-κB/Rel proteins and IκB-α in the nucleus (28–30). Moreover, neutrophil activation results in the parallel degradation of IκB-α in both the cytoplasm and nucleus (29, 30). More recently, we have shown that IKK phosphorylation, IκB-α phosphorylation/degradation, and RelA phosphoryation all occur in both cellular compartments of neutrophils (30). These observations prompted us to investigate the upstream kinase(s) that are responsible for the activation of the nuclear and cytoplasmic IKK complexes of human neutrophils. We now report that TAK1 is expressed in both the cytoplasm and the nucleus, and that it is rapidly and transiently activated by TNF-α or LPS. Inhibition of TAK1 mostly abrogates the inducible phosphorylation of IKKα/β and RelA, as well as the degradation of IκB-α, that occur in both the nuclear and cytoplasmic compartments. Accordingly, TAK1 inhibition also prevented inducible NF-κB binding to DNA and κB-driven promoter activation and strongly downregulated the κB-dependent transcription and secretion of inflammatory cytokines induced by TNF-α or LPS, as well as the delayed apoptosis elicited by these stimuli. Finally, we show that TAK1 also acts upstream of MAPKs known to influence both cytokine generation and delayed apoptosis in neutrophils. Our results therefore establish a crucial role for TAK1 in the activation of both nuclear and cytoplasmic signaling cascades in human neutrophils, as well as on downstream responses.
Materials and Methods
Abs and reagents
Abs raised against TAK1, TAB1, TAB3, MEKK1, MEKK3, all IKK isoforms, and IκB-α were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), whereas Abs to TAB2 and all phospho-Abs were from Cell Signaling Technology (Beverly, MA). The TAB4 Ab was a kind gift from Dr. D. Brautigan (University of Virginia, Charlottesville, VA). Ficoll-Paque, T4 polynucleotide kinase, polydeoxyinosinic-deoxycytidylic acid, and protein G-sepharose 4FF beads were from GE Healthcare Biosciences (Baie d’Urfé, Quebec, Canada). Radionucleotides were from NEN (Boston, MA). Endotoxin-free (<2 pg/ml) RPMI 1640 and FCS were from Wisent (St-Bruno, Quebec, Canada). Recombinant human cytokines were from R&D Systems (Minneapolis, MN), and Ultrapure LPS (from Escherichia coli 0111:B4) was from InvivoGen (San Diego, CA). Annexin V-FITC was from BD Biosciences (Missisauga, Ontario, Canada). Acetylated BSA, cycloheximide, diisopropyl fluorophosphate, myelin basic protein (MBP), PMSF, and propidium iodide were from Sigma-Aldrich (St. Louis, MO). The protease inhibitors, aprotinin, 4-(2-aminomethyl)benzenesulfonyl fluoride, leupeptin, and pepstatin A were all from Roche (Laval, Quebec, Canada). The TAK1 inhibitor, 5z-7-oxozeaenol, was a kind gift from Drs M. Tsuchiya and K. Ono (Chugai Pharmaceutical, Shizuoka, Japan). All other reagents were of the highest available grade, and all buffers and solutions were prepared using pyrogen-free clinical grade water.
Cell isolation and culture
Neutrophils were isolated from the peripheral blood of healthy donors from whom informed consent had been obtained under a protocol that was approved by an institutional ethics committee, as described previously (30). As determined by Wright staining and nonspecific esterase cytochemistry, the final neutrophil suspensions consistently contained <0.5% monocytes or lymphocytes, and neutrophil viability exceeded 98% after up to 4 h in culture, as determined by trypan blue exclusion. The myelomonoblastic PLB-985 cell line was purchased from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (Braunschweig, Germany). All cells were cultured at 37°C under a humidified 5% CO2 atmosphere in RPMI 1640 containing 10% FCS, 100 U/ml penicillin, and 100 μg/ml streptomycin (hereafter referred to as complete RPMI medium).
Denaturing electrophoreses and immunoblots
Cells were incubated at 37°C in the presence or absence of inhibitors and/or stimuli, as specified in the figure legends. Incubations were stopped by adding equivalent volumes of ice-cold PBS supplemented with diisopropyl fluorophosphate (2 mM, final concentration) and phosphatase inhibitors (10 mM NaF, 1 mM Na3VO4, and 10 mM Na4P2O7). Cells were then centrifuged and resuspended in relaxation buffer prior to disruption by nitrogen cavitation, as described previously (28, 31). Whole-cell samples or subcellular fractions were then prepared as described previously (28, 30, 31). Samples were electrophoresed on denaturing gels prepared according to the method of Laemmli (32); equal loading was ascertained by adjusting sample volumes based on their respective protein content. Following SDS-PAGE, proteins were transferred onto nitrocellulose membranes, which were stained with Ponceau Red, destained, and then processed for immunoblot analysis, as previously described (28).
Immunoprecipitations and immunoblots
Neutrophils (20 × 106 cells/1500 μl) were incubated at 37°C in the presence or absence of inhibitors and/or stimuli; reactions were terminated by a 15 s centrifugation at 12,000 × g. Cell pellets were lysed with 1 ml of ice-cold lysis buffer (20 mM Tris [pH 7.4], 1% Nonidet P-40, 5 mM EDTA, and 100 mM NaCl) containing protease and phosphatase inhibitors, and centrifuged at 15,000 rpm for 15 min at 4°C. Precleared lysates were incubated with 1.5 μg/ml mouse anti-TAK1 for 90 min at 4°C prior to an overnight incubation with protein G-sepharose 4FF beads at 4°C. Beads were washed twice in lysis buffer, then twice in PBS, and resuspended and boiled in 30 μl SDS sample buffer. The products were separated on Laemmli 10% gels, transferred onto nitrocellulose membranes, and then processed for immunoblot analysis with rabbit anti-IKKα/β, TAB1/2, or TAK1.
Immunoprecipitations and in vitro kinase assays
Neutrophils were stimulated, lysed, and immunoprecipitated as described above, using 1.5 μg/ml anti-TAK1, anti-MEKK1, or anti-MEKK3 Abs. Sepharose beads were washed once in lysis buffer, once in rinse buffer (20 mM Tris [pH 7.4], 150 mM NaCl, 10 mM Na4P2O7, 25 mM β-glycerophosphate, 1 mM Na3VO4, and 2 μg/ml aprotinin), and twice in PAN buffer (20 mM PIPES [pH 7], 2 μg/ml aprotinin, and 100 mM NaCl). Beads were then resuspended in 30 μl PAN buffer, from which 20 μl were added to 20 μl kinase buffer (20 mM PIPES [pH 7], 10 mM MgCl2, 2 mM DTT, 1 μg MBP, and 20 μM ATP) containing 5 μCi [γ-32P]-ATP (for TAK1 kinase assay), 20 mM PIPES (pH 7), 10 mM MnCl2, 2 mM DTT, 1 μg MBP, 20 μM ATP, and 5 μCi [γ-32P]-ATP (for MEKK1 kinase assay), or 20 mM PIPES (pH 7), 5 mM MnCl2, 5 mM MgCl2, 2 mM DTT, 1 μg MBP, 20 μM ATP, and 5 μCi [γ-32P]-ATP (for MEKK3 kinase assay). Samples were incubated (30 min, 30°C) with frequent mixing, and reactions were terminated by boiling in 5× sample buffer. The resulting samples were resolved on 18% SDS-PAGE, followed by drying and autoradiography. The remaining 10 μl resuspended beads (in PAN buffer) were boiled in 10 μl 1.5× sample buffer, resolved on 10% SDS-PAGE, transferred onto nitrocellulose membranes, and immunoblotted with Abs to TAK1, MEKK1, or MEKK3.
EMSAs
Isolation of RNA and real-time PCR analyses
Neutrophils were incubated in the presence or absence of stimuli or inhibitors for the desired times, as indicated. Total RNA was extracted following a slightly modified Chomczynski & Sacchi procedure (33) and reverse transcribed using random decamers (Ambion, Austin, TX) and SuperScript II (Invitrogen, Burlington, Ontario, Canada). The resulting cDNA was analyzed by semiquantitative real-time PCR assay following the procedure of Dussault and Pouliot (34), in a Rotorgene 3000 instrument (Corbett Research, Sydney, Australia). Oligonucleotide primers were as follows: IL-8 (forward 5′-AGGAAGCTCACTGGTGGCTG-3′; reverse 5′-TAGGCACAATCCAGGTG-GC-3′); MIP-1α (forward 5′-AGCTGACTACTTTGAGACGAGCA-3′; reverse 5′-CGGCTTCGCTTGGTTAGGA-3′); MIP-1β (forward 5′-CTGCT-CTCCAGCGCTCTCA-3′; reverse 5′-GTAAGAAAAGCAGCAGGCGG-3′); TNF-α (forward 5′-TCTTCTCGAACCCCGAGTGA-3′; reverse 5′-CC-TCTGATGGCACCACCAG-3′); and 18S (forward 5′-AGGAATTGACGGA-AGGGCAC-3′; reverse 5′-GGACATCTAAGGGCATCACA-3′). For primary transcript (PT)-PCR analyses, the following primers were used: IL-8 (forward 5′-ATTGAGAGTGGACCACACTG-3′; reverse 5′-ACTACTGTAATCCTA-ACACCTG-3′); MIP-1α (forward 5′-GTCAGTCCTTTCTTGGCTCTG-3′; reverse 5′-GATACCCACAACGAAACTCAGAC-3′); MIP-1β (forward 5′-AAACCTCTTTGCCACCAATACC-3′; reverse 5′-AGGACTCTGCCTACA-CCTTGAC-3′); and TNF-α (forward 5′-TCAGGATCATCTTCTCGAACC-3′; reverse 5′-GAGTCCTTCTCACATTGTCTC-3′).
Transfection of PLB-985 cells and luciferase assays
Granulocytic differentiation of PLB-985 cells was induced by including 1.25% DMSO in the culture medium, as described earlier (35). On the fifth day of differentiation, PLB-985 cells were nucleofected with pNFκB-Luc (Stratagene, La Jolla, CA), pIL8-Luc (a kind gift from Dr. Allan R. Brasier, University of Texas Medical Branch, Galveston, TX), or mock vectors, and luciferease activity was measured as described earlier (35).
ELISA analysis of secreted proteins
Neutrophils were cultured in 12-well culture plates at 37°C under a 5% CO2 atmosphere in the presence or absence of stimuli and/or inhibitors for the indicated times. Culture supernatants were collected, snap-frozen in liquid nitrogen, and stored at −70°C. Cytokine concentrations were determined in in-house sandwich ELISA assays, using commercially available capture and detection Ab pairs (R&D Systems; BD Biosciences).
Determination of neutrophil apoptosis
Neutrophils (5 × 105 cells) were washed twice in ice-cold PBS and incubated on ice for 30 min with FITC-conjugated Annexin V and, in some experiments, with propidium iodide. Stained cells were washed and analyzed (minimum of 10,000 cells) on a FACScan instrument (BD Biosciences) using the CellQuest software (BD Biosciences). Alternatively, caspase-3/7 activation was measured. For this purpose, neutrophils were plated and cultured overnight; 0.5 × 106 cells/condition were then processed for determination of caspase activation using the Caspase-Glo 3/7 according to the manufacturer’s intructions (Promega, Madison, WI) in a Berthold Instruments luminometer.
Results
Expression and distribution of TAK1 and TAB1/2 in resting and activated neutrophils
Our previous studies have shown that in human neutrophils, the three constituent subunits of the IKK complex (IKKα, IKKβ, and IKKγ) are expressed in both the cytoplasmic and nuclear compartments, where they can be activated in parallel (30). This prompted us to examine whether potential upstream mediators of IKKs are similarly localized. We therefore determined whether a promising candidate, TAK1, is expressed in neutrophils and where it might distribute within the cells. To this end, resting neutrophils were disrupted by nitrogen cavitation, and cell fractions were prepared. For comparison, autologous PBMCs were similarly prepared and analyzed in parallel. Fig. 1A shows that TAK1 and its binding partners, TAB1/2, distribute to both the cytoplasmic and nuclear compartments in neutrophils, whereas they are strictly cytoplasmic in autologous PBMCs. Another known binding partner of TAK1, TAB3, was not detected in neutrophils (not shown). The unexpected detection of TAK1 and TAB1/2 in the nucleus of neutrophils prompted us to ascertain the purity of our subcellular fractions, which were analyzed for the presence of cytosolic and nuclear markers. As shown in Fig. 1B, strictly cytosolic proteins, such as lactate dehydrogenase and leukotriene A4 hydrolase (36), were only detected in cytoplasmic fractions. Conversely, histone H3 was exclusively nuclear, as expected (Fig. 1B). Thus, our nuclear and cytoplasmic fractions are reasonably exempt from cross-contamination. This is consistent with our previous studies using similar or identical cell fractionation procedures (28, 30).
We next examined the fate of TAK1 following neutrophil activation by physiological stimuli. Exposure of the cells to NF-κB activators, such as TNF-α or LPS, did not affect the subcellular distribution of TAK1 for up to 4 h (Fig. 1C and data not shown). The cellular levels of TAK1 were also unaffected by cycloheximide for up to 4 h, whether in resting or activated cells, indicating a slow turnover rate (Fig. 1D); similar results were observed in the case of TAB1 and TAB2 (data not shown). By comparison, IκB-α was rapidly degraded in unstimulated cells when protein synthesis was blocked, and its de novo synthesis following TNF-α or LPS stimulation was also prevented by cycloheximide, as expected (Fig. 1D). Thus, IκB-α turnover was rapid and dependent on protein synthesis, as we previously reported in neutrophils (28, 30), whereas cellular levels of TAK1, TAB1, and TAB2 were found to be very stable. Finally, blocking nuclear export with leptomycin B had no effect on the cellular distribution of TAK1, even after 6 h (Supplemental Fig. 1). Thus, TAK1 does not seem to shuttle between the nucleus and cytoplasm in neutrophils.
Association among TAK1, TAB1, TAB2, TAB4, and IKKα/β in human neutrophils
It has been shown that TAK1 can associate with, and/or dissociate from, TAB1/2 or IKKα/β following cell stimulation and that these various associations participate in the regulation of IKK activity (12, 17, 20, 21). To determine whether these phenomena also take place in primary neutrophils, the cells were stimulated with TNF-α or LPS for various times, and the resulting cell lysates were immunoprecipitated with mouse anti-TAK1 and immunoblotted with rabbit Abs to TAB1, TAB2, IKKα/β, or (for loading control) to TAK1 itself (using a different Ab than the one used for immunoprecipitation). As shown in Fig. 2, TAK1 is constitutively associated with TAB1/2, as well as with IKKα/β, in unstimulated neutrophils. No changes in these interactions occurred following cell stimulation with either TNF-α or LPS (Fig. 2). Identical results were obtained when the cell lysates were immunoprecipitated with anti-IKKα/β and revealed with anti-TAK1 Abs (data not shown). This suggests that in human neutrophils, there is no need for an inducible recruitment of TAK1 to a multimeric complex containing IKK subunits and/or TAB1/2 isoforms, as observed in some cellular settings (12, 17), because TAK1 is constitutively associated with all these signaling partners.
Other signaling partners that are important for IKK activity have been reported to associate with (or dissociate from) a TAK1/IKK holocomplex. These notably include the newly characterized type 2A phosphatase binding protein TAB4 (37). We therefore determined whether TAB4 can be detected in IKK or TAK1 immunoprecipitates from resting and activated neutrophils. As shown in Fig. 2C, TAB4 is constitutively associated with the IKK complex, and this association is unaffected in neutrophils stimulated with either LPS or TNF-α; similar results were observed when neutrophils were stimulated for up to 2 h (not shown). Thus, TAB4 is yet another constituent of the TAK1/IKK complex of neutrophils.
Activation and inhibition of TAK1 in human neutrophils
We next investigated whether TAK1 becomes activated in response to neutrophil stimulation. For this purpose, cells were stimulated with TNF-α or LPS over a short time interval, as MAP3Ks are often activated at very early time points. The corresponding cell lysates were then immunoprecipitated using an anti-TAK1 Ab, and TAK1 activity was determined in in vitro kinase assays using MBP as a substrate. As shown in Fig. 3A, the phosphorylation of MBP in TAK1 immunoprecipitates was rapidly and transiently elicited in response to TNF-α or LPS, peaking by 5 min and rapidly declining thereafter. Thus, TAK1 is activated with rapid kinetics in human neutrophils.
To delineate the role of TAK1 activation in downstream processes, we sought to inihibit the kinase. In this regard, 5z-7-oxozeaenol was reported to be a highly selective inhibitor of TAK1, in that it blocked TAK1 activity without significantly affecting the activities of other MAP3Ks, such as MEKK1 or apoptosis signal-regulating kinase (ASK)-1 (38). To determine the 5z-7-oxozeaenol concentrations that effectively inhibit TAK1 activity in neutrophils, cells were preincubated for 30 min with increasing concentrations of the inhibitor prior to stimulation with TNF-α; the corresponding lysates were then immunoprecipiated using Abs to TAK1 or other MAP3Ks present in neutrophils and processed for kinase assays. As shown in Fig. 3B, TAK1 kinase activity induced by TNF-α was progressively attenuated by increasing concentrations of the TAK1 inhibitor, an almost complete inhibition being consistently achieved using 1 μM 5z-7-oxozeaenol. By contrast, this inhibitor (whether at 1 μM or higher concentrations) failed to affect MEKK1 or MEKK3 kinase activities in neutrophils (Fig. 3C, 3D). Similar results were obtained when neutrophils were stimulated with LPS instead of TNF-α (data not shown). These results therefore establish that 5z-7-oxozeaenol inhibits TAK1 in a highly selective manner in primary human neutrophils.
Impact of TAK1 on the IKK/NF-κB cascade of human neutrophils
Our previous studies demonstrated that in neutrophils, the phosphorylation of IKK isoforms (α/β/γ) induced by TNF-α or LPS occurs in both the nucleus and cytoplasm (30). To determine whether TAK1 represents an upstream IKK kinase, neutrophils were pretreated with increasing concentrations of 5z-7-oxozeaenol prestimulation with TNF-α or LPS; the cytoplasmic and nuclear fractions were then subjected to immunoblot analysis. As shown in Fig. 4A and 4B, 1 μM 5z-7-oxozeaenol (i.e., a concentration that was found to block TAK1 kinase activity in these cells) almost completely abrogated the inducible phosphorylation of IKKα/β and RelA and prevented the inducible degradation of IκBα in both the nuclear and cytoplasmic fractions of neutrophils. Accordingly, inducible NF-κB DNA binding was mostly abolished by 1 μM 5z-7-oxozeaenol in neutrophils (Fig. 4C), and the activity of κB-driven promoters was similarly abrogated by 1 μM 5z-7-oxozeaenol in human granulocytic PLB-985 cells (Supplemental Fig. 2). By contrast, the same concentration of 5z-7-oxozeaenol had no effect on the LPS-elicited phosphorylation of IKKγ (Fig. 4B) and only marginally affected IKKγ phosphorylation induced by TNF-α (Fig. 4A). These results clearly establish that in human neutrophils, TAK1 is a key upstream regulator of most activation events associated with the cytoplasmic and nuclear IKK signaling cascade, as well as with NF-κB DNA binding and subunit phosphorylation.
Effect of TAK1 inhibition on inflammatory cytokine expression and release in human neutrophils
We recently showed that a number of inflammatory cytokines generated by activated neutrophils are under the control of NF-κB (39). We therefore investigated the consequence of TAK1 inhibition on the ability of neutrophils to produce such inflammatory mediators. Cells were pretreated with 1 μM 5z-7-oxozeaenol prior to stimulation with TNF-α or LPS, and cytokine generation was then assessed. As shown in Fig. 5A, TAK1 inhibition strongly repressed the inducible gene expression of all inflammatory mediators investigated (namely, IL-8, TNF-α, MIP-1α, and MIP-1β). Further investigation showed that the TAK1 inhibitor exerted a comparable effect on the inducible transcription of the same genes in neutrophils, as assessed in PT-PCR analyses (Fig. 5B). This indicates that the effect of 5z-7-oxozeaenol on steady-state mRNA levels most likely reflects transcriptional inhibition. In support of this conclusion, we observed that 5z-7-oxozeaenol effectively inhibited the transactivation of a κB-driven promoter, and of the full-length IL-8 promoter, in human neutrophil-like PLB-985 cells stimulated with either LPS or TNF-α (Supplemental Fig. 2). Finally, and in agreement with the above data, the inducible secretion of the corresponding proteins was also profoundly decreased in neutrophils pretreated with 5z-7-oxozeaenol (Fig. 6). Thus, TAK1 emerges as a crucial signaling molecule for the generation of inflammatory cytokines in neutrophils activated by physiological agonists.
Effect of TAK1 inhibition on other signaling cascades affecting inflammatory cytokine generation in human neutrophils
We recently demonstrated the involvement of the p38 MAPK and MEK/ERK pathways in the production of inflammatory cytokines by activated neutrophils (39). Because TAK1 has been reported to act upstream of both p38 MAPK and ERK in some cellular settings (15, 40), we investigated whether TAK1 inhibition might affect these MAPK pathways in neutrophils. As shown in Fig. 7A, cell pretreatment with 5z-7-oxozeaenol prevented the LPS- and TNF-elicited phosphorylation of p38 MAPK and ERK. Similar results were observed in the case of the JNK cascade (Supplemental Fig. 4). Thus, TAK1 acts upstream of the main MAPK pathways in human neutrophils.
We also investigated the effect of TAK1 inhibition on IFN-γ signaling, because the latter is required as a costimulus for the production of several κB-dependent cytokines in neutrophils (including CXCL9, CXCL10, CXCL11, and IL-12) (41–43). For this purpose, we examined the inducible phosphorylation of STAT1, which occurs in response to IFN-γ in neutrophils (31, 44). As shown in Fig. 7B, inhibition of TAK1 had no effect on the phosphorylation of STAT1 induced by IFN-γ, in keeping with the fact that IFN-γ has not been reported to signal through TAK1. In the same experiments, the cellular levels of IκB-α were also unaffected by the TAK1 inhibitor or by IFN-γ stimulation, consistent with the fact that the cytokine fails to induce NF-κB activation or IκB-α degradation in human neutrophils (28, 45). Together, the above results demonstrate a selective involvement of TAK1 in the various signaling pathways mobilized in activated neutrophils.
Effect of TAK1 inhibition on spontaneous and delayed apoptosis in human neutrophils
Neutrophils readily become apoptotic within several hours of culture, a phenomenon that can, however, be delayed following exposure to most physiological neutrophil agonists. In the case of stimuli, such as LPS and TNF-α, the apoptosis-delaying effect has been reported to involve the NF-κB as well as p38 MAPK and ERK pathways (46–51). Because of our finding that TAK1 acts upstream of all three signaling pathways in neutrophils, we investigated whether TAK1 inhibition might interfere with neutrophil survival induced by LPS or TNF-α. As a control, we also examined the effect of TAK1 inhibition on the antiapoptotic effect of dexamethasone, a stimulus that does not use the NF-κB, p38 MAPK, or ERK pathways. As shown in Fig. 8, constitutive neutrophil apoptosis was not affected by the presence of 5z-7-oxozeaenol in the culture medium, showing that the inhibitor does not accelerate (or delay) apoptosis on its own. Likewise, the TAK1 inhibitor did not promote neutrophil necrosis (Supplemental Fig. 3). By comparison, TAK1 inhibition reversed the antiapoptotic effect of LPS and TNF-α, but failed to affect the dexamethasone-elicited neutrophil survival, as expected (Fig. 8B). Thus, TAK1 appears to be a crucial intermediate governing the modulation of neutrophil apoptosis in response to stimuli that use the NF-κB and/or MAPK signaling cascades.
Discussion
Neutrophils are key players in innate immunity and can also influence adaptive responses (27). These biological actions of neutrophils largely depend on their ability to generate cytokines, a functional response that is under the control of the IKK/NF-κB and MAPK signaling cascades (35, 39, 52–55). In this study, we report for the first time that the MAP3K, TAK1, is expressed and activated by inflammatory stimuli in human neutrophils and that it is crucially involved in several signaling pathways and downstream functional responses. We found that TAK1 and its associated proteins, TAB1, TAB2, and TAB4, are all expressed in both the cytoplasmic and nuclear compartments of neutrophils, in contrast to their strictly cytoplasmic localization in other cellular systems (as shown in this paper in the case of human PBMCs). This is reminiscent of our previous finding that the three major IKK isoforms are also expressed (and activated) in both the nucleus and cytoplasm of neutrophils (30), whereas they are typically cytoplasmic in other cell types. Moreover, we showed that in neutrophils, TAK1, TAB1, TAB2, and TAB4 are constitutively associated with IKK subunits and that this association is unchanged following cell stimulation. Thus, it appears that TAK1 and the TAB proteins form a stable complex with the IKK signalosome in neutrophils, both in the nucleus and cytoplasm. Such a high m.w. complex would not be expected to traverse the nuclear pore, and, accordingly, we found that neither TAK1 nor individual IKK isoforms shuttle between the nuclear and cytoplasmic compartments in neutrophils (30 and this paper). Whereas the constitutive and persistent association of TAB1 with TAK1 (observed in this study in neutrophils) is in keeping with the existing literature, the constitutive association of TAK1 with TAB2 and TAB4 could represent yet another distinctive feature of neutrophil signaling, insofar as these two isoforms were reported to coimunoprecipitate with TAK1 only after cell stimulation in other systems (20, 37).
The presence of TAK1, TAB1, and TAB2 within a common complex in resting neutrophils meets the basic requirements for a prompt activation of TAK1 upon cell activation (21), and we indeed found that TAK1 kinase activity was induced within minutes in response to LPS or TNF-α in neutrophils. Another contributing factor might be the seemingly permanent association of TAK1 with TAB4, as the latter was recently shown to directly bind to TAK1 and to activate it to an extent that exceeded that achieved by TAB1 (37). Interestingly, TAB4 can also bind the phosphatase PP6 (37, 56), which is known to dephosphorylate TAK1 (57), and it was proposed that polyubiquitin scaffolds and PP6 might be mutually exclusive TAB4 binding partners (37). Thus, TAB4 could serve a dual purpose in the context of TAK1 activation, and it is tempting to speculate that the persistent association of TAB4 with TAK1 observed in neutrophils might promote the recruitment of phosphatases at later time points. This would help explain the highly transient nature of TAK1 activity in neutrophils, which had mostly subsided by 10 min of stimulation. Alternatively, it was recently reported that the F-box protein FBXW5 can inducibly associate with TAK1, resulting in its dephosphorylation (58). Whatever the case may be, the termination of TAK1 activity remains a poorly understood issue and stresses the need for further investigation. A similar conclusion can be drawn in the case of the mechanisms underlying TAK1 activation, which also remain unclear. Some investigators have proposed the existence of a TAK1 kinase, but a candidate has yet to be pinpointed. Alternatively, there exists evidence that TAK1 can be activated by autophosphorylation (or transautophosphorylation as TAK1-containing complexes oligomerize following cell stimulation) in a process that is seemingly associated with the ubiquitination state of the kinase, or at least its interaction with polyubiquitin chains (14). A recent study even showed that unanchored polyubiquitinated chains can cause TAK1 to autophosphorylate in an in vitro assay involving recombinant proteins (59). Again, further studies are needed to elucidate this aspect of TAK1 regulation.
Our observation that the kinetics of TAK1 activation in neutrophils preceded IKK subunit phosphorylation and IκB-α degradation in both the nuclear and cytoplasmic fractions suggested that TAK1 might function as an IKK kinase in both cellular compartments. To address this possibility, we investigated the consequences of pretreating neutrophils with the highly selective TAK1 inhibitor 5z-7-oxozeaenol (38) on various steps of the NF-κB pathway in these cells. Pharmacological inhibition of TAK1 revealed that the kinase is involved in most aspects of the NF-κB pathway in neutrophils. Indeed, 5z-7-oxozeaenol prevented the LPS- or TNF-elicited phosphorylation of IKKα/β and RelA, as well as the degradation of IκB-α, in both the cytoplasmic and nuclear compartments. This was paralleled by a profound inhibition of downstream events, such as nuclear NF-κB DNA binding and κB-driven promoter transactivation. Accordingly, the gene transcription, steady-state gene expression, and release of several cytokines (i.e., TNF-α, IL-8, MIP-1α, and MIP-1β), which are all known to depend on NF-κB in neutrophils (39), were also mostly abrogated by the TAK1 inhibitor in these cells. We additionally showed that TAK1 inhibition prevents the activation of the p38 MAPK and ERK signaling pathways, which are known to affect neutrophil cytokine generation independently of the IKK/NF-κB cascade (39). Thus, the regulation of neutrophil cytokine production by TAK1 is multipronged. We further showed that TAK1 acts upstream of JNK, though the latter does not appear to contribute much to cytokine generation in neutrophils (60). In good agreement with the above results, we found that TAK1 affects other neutrophil responses involving the IKK/NF-κB and MAPK signaling cascades. One such response is the apoptosis-delaying effect of LPS or TNF-α (46–51), which we found to be abolished in cells treated with the TAK1 inhibitor. In addition, JNK was suggested to influence neutrophil apoptosis (61) and might therefore contribute to the effect of TAK1 on this response, given that JNK phosphorylation depends on TAK1 in these cells, as shown in this study. Together, our findings demonstrate that in primary human neutrophils, TAK1 activation is a critical upstream event for the mobilization of the nuclear and cytoplasmic IKK/NF-κB cascade, and of MAPK pathways, which control key functional responses of these cells.
Even though 5z-7-oxozeaenol was previously shown to be highly selective for TAK1 among various MAP3Ks (38), its effects in neutrophils were often very pronounced, an outcome that urges caution, especially because our conclusions on the pivotal role of TAK1 in neutrophil responses largely rely on the use of this inhibitor. In this respect, a number of observations demonstrate that the TAK1 inhibitor does indeed act in a highly selective manner in human neutrophils. First, 5z-7-oxozeaenol did not prove to be a general inhibitor of neutrophil function, as it had no effect on processes, such as spontaneous apoptosis. Similarly, the inhibitor did not alter the cycle threshold of GAPDH or 18S rRNA when samples were analyzed in quantitative PCR (data not shown), showing that it does not block transcription nonspecifically. Secondly, the TAK1 inhibitor failed to affect signaling cascades that are independent of TAK1, such as the IFN-γ–elicited phosphorylation of STAT1 or the dexamethasone-elicited enhancement of neutrophil survival (which does not depend on NF-κB or MAPKs). Thirdly, even though 5z-7-oxozeaenol profoundly inhibited several responses elicited by LPS or TNF, it did not block the inducible phosphorylation of IKKγ or the inducible activation of MEKK1 or MEKK3 occurring under these conditions. Thus, the TAK1 inhibitor did not behave as a general inhibitor of LPS or TNF signaling in neutrophils. More importantly, our observation that 5z-7-oxozeaenol does not affect MEKK1 and MEKK3 activity strongly indicates that the results we obtained using the inhibitor cannot be attributed to a nonselective inhibition of other MAP3Ks expressed in neutrophils. This is especially significant in view of the fact that both MEKK1 and MEKK3 were suggested to act as IKK kinases in other experimental settings (7, 62–64). This being said, it cannot be excluded that MEKK1 and MEKK3 might participate in IKK activation. Although our data show that TAK1 does not act upstream of the MEKK isoforms, the latter might act upstream of TAK1 or be activated in parallel. In this regard, MEKK3 was reported to cooperate with TAK1 in inducing NF-κB activation in some instances, with TAK1 phosphorylating IKKβ and MEKK3 phosphorylating IKKγ (9). Whether MEKK3 accounts for the TAK1-independent phosphorylation of IKKγ observed in neutrophils, however, remains to be demonstrated. Incidentally, our results represent the first demonstration that MEKK3 is expressed and activatable in human neutrophils and also show for the first time that MEKK1 activity can be triggered in response to LPS, as previously reported for fMLP and (in adherent cells) for TNF (61, 65).
In conclusion, our findings significantly extend our understanding of the signaling machinery upstream of IκB-α and MAPKs in primary human neutrophils (summarized in Fig. 9) (66) and further emphasize the unique characteristics of the IKK cascade in these cells by showing that TAK1 acts as an IKK kinase in the nucleus as well as in the cytoplasm. This therefore makes TAK1 an attractive target for therapeutic intervention in view of the foremost role played by neutrophils in several chronic inflammatory conditions.
Acknowledgements
Disclosures The authors have no financial conflicts of interest.
Footnotes
This work was supported by grants to P.P.M. from the Canadian Arthritis Society and the Canadian Institutes of Health Research. C.F.F. holds a Ph.D. Scholarship from the Canadian Institutes of Health Research.
The online version of this article contains supplemental material.