The survival effect of TNF-α  in human neutrophils is mediated via NF-κB-dependent IL-8 release

Authors

  • Andrew S. Cowburn,

    Corresponding author
    1. Respiratory Medicine Division, Department of Medicine, University of Cambridge School of Clinical Medicine, Addenbrooke's and Papworth Hospitals, Cambridge, GB
    • Respiratory Medicine Division, Department of Medicine, Level 5, Box 157, Addenbrooke's Hospital, Hills Road, Cambridge, CB2 2QQ, GB Fax: +44-1223-762007
    Search for more papers by this author
  • John Deighton,

    1. Respiratory Medicine Division, Department of Medicine, University of Cambridge School of Clinical Medicine, Addenbrooke's and Papworth Hospitals, Cambridge, GB
    Search for more papers by this author
  • Sarah R. Walmsley,

    1. Respiratory Medicine Division, Department of Medicine, University of Cambridge School of Clinical Medicine, Addenbrooke's and Papworth Hospitals, Cambridge, GB
    Search for more papers by this author
  • Edwin R. Chilvers

    1. Respiratory Medicine Division, Department of Medicine, University of Cambridge School of Clinical Medicine, Addenbrooke's and Papworth Hospitals, Cambridge, GB
    Search for more papers by this author

Abstract

The capacity of cytokines to modulate neutrophil apoptosis is thought to be a major factor influencing the resolution of granulocytic inflammation. We have previously shown that the late survival effect of TNF-α  in human neutrophils involves activation of both NF-κB and phosphoinositide 3-kinase (PI3-kinase) pathways. In this study, we address how these pathways integrate to prevent cell death. In human neutrophils, TNF-α  (200 U/ml) induced rapid IκB-α  degradation, NF-κB activation and IL-8 release (31.8±5.4 pg/105 cells/2 h), whereas GM-CSF (10 ng/ml) stimulated an equivalent IL-8 release (26.5±4.5 pg/105 cells/2 h) without enhanced IκB-α  degradation or NF-κB activation compared to control. Importantly, inhibition of PI3-kinase did not modify TNF-α -induced IκB-α  degradation, yet fully inhibited the survival effect of both cytokines. Inhibition of IκB-α  phosphorylation, PI3-kinase or ERK1/2 activation blocked IL-8 release by both cytokines. Blocking IL-8 activity by inhibiting its synthesis or by using a neutralizing antibody enhanced the early pro-apoptotic effectof TNF-α  and inhibited its late survival effect without affecting GM-CSF-induced survival. These data suggest that cross-talk between NF-κB and PI3-kinase pathways in TNF-α -stimulated neutrophils results from NF-κB/ERK1/2-dependent IL-8 production which acts in an autocrine manner to drive PI3-kinase-dependent survival. In contrast, GM-CSF-mediated survival does not involve NF-κB activation or IL-8 release.

Abbreviations:
BAY11–7085:

Butylphenyl-sulfonyl-propenenitrile

fMLP:

N-Formyl-Met-Leu-Phe

IκK:

IκB kinase

Parthenolide:

Sesquiterpene-lactone

PI3-kinase:

Phosphoinositide 3-kinase

PtdIns(3,4,5)P3:

Phosphatidylinositol 3,4,5-trisphosphate

PtdIns(4,5)P2:

Phosphatidylinositol 4,5-bisphosphate

1 Introduction

Apoptotic cell death has been proposed as a central mechanism underlying the ordered elimination of neutrophils from an inflammatory site and, hence, to play a key role in the resolution of inflammation 1, 2. Dysregulation of this process by various inflammatory mediators may contribute directly to the pathogenesis of several diseases characterized by neutrophil-mediated tissue injury, including acute respiratory distress syndrome (ARDS) 3, 4. We and others have identified the importance of both the phosphoinositide 3-kinase (PI3-kinase) and NF-κB pathways in modulating the rate of neutrophil apoptosis 57. However, little data exists examining the potential interactions between these two pathways in primary cells and, in particular, the role of NF-κB in PI3-kinase-dependent neutrophil survival. Cytokine-stimulated accumulation of phosphatidylinositol 3,4,5-trisphosphate [PtdIns(3,4,5)P3] 8 activates a number of downstream targets including Akt 9, 10. Akt in turn phosphorylates effector proteins, among these the pro-apoptotic factor Bad, pro-caspase-9 and the forkhead transcription factors 1115.

NF-κB is also a key transcriptional activator of several anti-apoptotic proteins including Bcl-2, Bcl-X and IL-8 1619. The activation of NF-κB is regulated by a complex family of proteins termed IκBα /β /ϵ  that sequester the transcription factor in a non-activated state within the cytosol 20, 21. Phosphorylation of the IκB complex by IκB kinase (IκK) leads to its poly-ubiquitination and subsequent degradation, enabling NF-κB to translocate to the nucleus 22, 23. In addition to this classic route of NF-κB activation, there appears to be a second regulatory level, independent of IκB degradation, that relies on the phosphorylation of serine residues in the transactivation domain of the p65 subunit of NF-κB 24, 25.

Several laboratories have recently proposed that various growth factors and cytokines may require PI3-kinase/Akt signaling for full NF-κB activation 2628. However, the exact mechanism and universal nature of PI3-kinase/Akt and NF-κB pathway interactions remains controversial 29, not least because many of these data were generated using transformed cells and overexpression systems. While activation of the Akt pathway has been reported to stimulate IκK-dependent IκB degradation and nuclear transactivation of NF-κB 13, 14, others have concluded that Akt-dependent activation of NF-κB occurs predominantly by stimulating phosphorylation of the transactivation domain of the p65 subunit rather than by inducing the degradation of IκB 24, 27. Moreover, phosphatase and tensin homologue (PTEN), a lipid phosphatase that converts PtdIns(3,4,5)P3 to phosphatidylinositol 4,5-bisphosphate [PtdIns(4,5)P2] and thereby reduces Akt activity, has been shown to inhibit TNF-α -stimulated NF-κB-dependent transcription 26 without affecting IκBα  degradation, p105 processing, p65 (RelA) nuclear translocation or DNA binding of NF-κB 26.

The aim of this study was to determine the degree and mechanism of cross-talk between the PI3-kinase and NF-κB pathways in a well characterized primary cell model, namely the neutrophil, in which independent activation of either pathway produces a marked survival response 5, 6. In particular, we wished to determine the role of NF-κB pathway activation in GM-CSF-mediated neutrophil survival and to explain the PI3-kinase dependence of the TNF-α -induced survival effect, since this cytokine has no direct effect on PtdIns(3,4,5)P3 accumulation or Akt activity. We show that GM-CSF enhances the NF-κB-dependent secretion of IL-8 without stimulating IκB degradation or mRNA transcription and without enhancing NF-κB activity. Of note, the enhanced release of IL-8 does not contribute to the survival effect of GM-CSF. In contrast, TNF-α -stimulated neutrophil survival is entirely IL-8 dependent (as indicated by the complete ablation of the survival signal following IL-8 neutralization), and this autocrine effect accounts for the PI3-kinase dependency of TNF-α -mediated survival. This study has identified the level and extent of PI3-kinase and NF-κB interaction in a non-manipulated primary cell system and has revealed the mechanism underlying the PI3-kinase-dependent survival effect of TNF-α  in neutrophils.

2 Results

2.1 Effect of inflammatory cytokines on the rate of constitutive neutrophil apoptosis

GM-CSF (10 ng/ml) inhibited apoptosis at both 6 and 20 h (% apoptosis; control 6 h 10.7±0.9%, GM-CSF 6 h 4.1±0.4%, p<0.005; control 20 h 72.3±3.2%, GM-CSF 20 h 11.9±2.3%, p<0.005, n=8; Fig. 1). In contrast, TNF-α  (200 U/ml) had a biphasic effect on apoptosis with an early (6 h) enhancement of the rate of apoptosis (% apoptosis; control 6 h 10.7±0.9%, TNF-α  6 h 32.3±3.7, p<0.005, n=8) and a later (20 h) inhibition (% apoptosis, control 20 h 72.3±3.2%, TNF-α  20 h 42.1±4.2%, p<0.005, n=8; Fig. 1). While the PI3-kinase inhibitor LY294002 did not influence the 6-h pro-apoptotic effect of TNF-α , the 20-h survival effect was attenuated by 98.5±13.8% (p<0.05, n=8; Fig. 1). The cytoprotective effect of GM-CSF at 20 h was also significantly attenuated by pre-incubation with LY294002 (10 μ M) (64.1±8.4%, p<0.005; Fig. 1). Of note, LY294002 alone had no effect on the rate of constitutive neutrophil apoptosis at either 6 or 20 h.

Figure 1.

GM-CSF and TNF-α  promote neutrophil survival by a PI3-kinase-dependent mechanism. Peripheral blood neutrophils were cultured in the presence or absence of GM-CSF (10 ng/ml) or TNF-α  (200 U/ml) with or without the PI3-kinase inhibitor LY294002 for the indicated time periods. Cells were harvested after 6 or 20 h, and cytospins were fixed and stained as detailed in Sect. 4.2. GM-CSF significantly inhibited the constitutive rate of neutrophil apoptosis at 6 and 20 h; in contrast, TNF-α  enhanced the rate of apoptosis at 6 h and delayed apoptosis at 20 h. Data represent percentages of apoptosis from eight independent experiments, each performed in triplicate, together with mean ± SEM values. *p>0.05.

2.2 The role of the NF-κB pathway in GM-CSF- and TNF-α -stimulated neutrophil survival

To examine the role of the NF-κB pathway in mediating the anti-apoptotic effect of GM-CSF and TNF-α , we initially assessed the effects of two independent and site-specific NF-κB inhibitors. Butylphenyl-sulfonyl-propenenitrile (BAY11–7085), which was previously shown to inhibit TNF-α -stimulated IκB degradation in epithelial cells 30, 31, had no effect on GM-CSF-stimulated neutrophil survival at either 6 or 20 h (Fig. 2A) when used at concentrations (<10 μ M) that did not affect the intrinsic rate of apoptosis. In contrast, BAY11–7085 over the same concentration range caused a significant increase in early TNF-α -stimulated apoptosis (% apoptosis at 6 h, control 8.7±0.5%, TNF-α  21.2±2.1%, BAY11-7085 [3 μ M] 13.1±1.2%, TNF-α / BAY11–7085 37.7±10.7%, p<0.05; data not shown) and inhibited TNF-α -stimulated survival at 20 h (% apoptosis at 20 h, control 61.9±0.7%, TNF-α  51.2±0.4%, BAY11–7085 [3 μ M] 66.7±2.8%, TNFα /BAY11–7085 63.6±1.9%, p<0.05; Fig. 2A–C). Sesquiterpene-lactone (parthenolide), which was previously shown to inhibit the phosphorylation of the transactivation domain of the p65 subunit and subsequent DNA binding in murine lung tissue and HeLa cells 32, 33, produced very similar data to those obtained with BAY11–7085 (Fig. 2D–E). These data suggest that the NF-κB pathway is tonically activated in cultured neutrophils and plays an important role in determining the rate of constitutive and TNF-α -, but not GM-CSF-regulated apoptosis.

Figure 2.

GM-CSF delays neutrophil apoptosis via a NF-κB-independent pathway. Human neutrophils were incubated in the presence or absence of GM-CSF (10 ng/ml) or TNF-α  (200 U/ml) with or without BAY11– 7085 (concentration response curve 0.1–20 μ M) for 20 h (A). Cells were harvested and stained with annexin V-FITC (a marker of phosphatidylserine exposure/apoptosis) and propidium iodide (a marker of plasma membrane integrity/post-apoptotic necrosis) as detailed in Sect. 4.3. Data represent values from three independent experiments, each performed in triplicate, together with mean ± SEM values. (B–E) Representative fluorescent histograms of 10,000 events showing the level of annexin V-FITC binding at 20 h for (B) control/BAY11–7085, (C) TNF-α  /TNF-α  + BAY11– 7085, (D) control/parthenolide and (E) TNF-α /TNF-α  + parthenolide-treated cells. TNF-α  (20 h) partially inhibited annexin V-FITC staining, indicative of an inhibition of neutrophil apoptosis. The survival effect was abolished by co-incubation with NF-κB inhibitors. Statistical analysis of controls via paired t-test: * or # p>0.05. Statistical analysis of concentration response curve by two-way ANOVA test compared to matched controls: # p>0.05 (analysis of TNF-α  against TNF-α /BAY11–7085).

2.3 Effect of PI3-kinase inhibition on cytokine-dependent IκB degradation

To examine the potential role of PI3-kinase in the phosphorylation and subsequent degradation of IκB, human neutrophils were pre-incubated with LY294002 (10 μ M) for 20 min prior to stimulation with GM-CSF (10 ng/ml) or TNF-α  (200 U/ml). Irrespective of PI3-kinase inhibition, TNF-α  (10 min) stimulated a marked decrease in IκB protein present in neutrophils. In contrast, the treatment of neutrophils with GM-CSF for 10 min (Fig. 3A) or 30 min (data not shown) had no effect on IκB stability. Likewise, analysis of IκB mRNA (an important transcriptional target for NF-κB) by semi-quantitative RT-PCR at 120 min identified a fourfold increase in TNF-α -treated neutrophils but again no effect by GM-CSF or LY294002 (Fig. 3B). NF-κB activity was determined by an ELISA-based assay detecting the p50 and p65 subunits in the nuclear fraction of cell lysates. TNF-α  stimulated the translocation of both subunits to the nuclear fraction; in contrast, GM-CSF had no effect on NF-κB activity as determined in this assay (Fig. 3C). The efficacy of LY294002 to inhibit PI3-kinase was shown in parallel assays examining PI3-kinase activity and N-formyl-Met-Leu-Phe (fMLP)-stimulated superoxide anion generation (data not shown). These data indicate that despite its ability to activate PI3-kinase 6, GM-CSF has no effect on IκB degradation, IκB gene transcription or NF-κB activation.

Figure 3.

TNF-α  stimulates NF-κB activation and IκB degradation via a PI3-kinase-independent pathway. Human neutrophils were incubated in the presence or absence of GM-CSF (10 ng/ml) or TNF-α  (200 U/ml) with or without LY294002 (10 μ M) at 37°C. Neutrophil lysates were prepared and IκB was quantified by Western blotting and RT-PCR. NF-κB (p50, p65) activity was quantified by an ELISA-based assay as detailed in Sect. 4.5. (A) Representative immunoblots (with corresponding densitometry) demonstrating TNF-α -stimulated rapid degradation of IκB via a PI3-kinase-independent mechanism. GM-CSF is unable to stimulate IκB degradation at 10 min. Identical data were obtained in three further independent experiments. (B) IκB is transcriptionally regulated by NF-κB. TNF-α  stimulates a fourfold increase in IκB  transcription (120 min) via a PI3-kinase-independent mechanism. The rate of IκB transcription in GM-CSF-stimulated cells does not deviate from that of the control. (C) The association of the p50 and p65 subunits with the nuclear fraction of cell lysates signifies NF-κB activity. TNF-α  stimulated the translocation of both NF-κB subunits to the nuclear fraction; in contrast, GM-CSF-stimulated NF-κB translocation did not deviate from that observed under control conditions (n=5). *p>0.05.

2.4 The role of PI3-kinase/ERK1/2 in GM-CSF- and TNF-α -stimulated IL-8 production

We initially confirmed that both TNF-α  34, 35 and GM-CSF 35, 36 stimulate IL-8 release from human neutrophils (Fig. 4). To investigate whether GM-CSF stimulates IL-8 synthesis via a NF-κB-dependent pathway, IL-8 release was assayed in supernatants obtained from cells incubated with or without GM-CSF or TNF-α  in the presence and absence of selective NF-κB inhibitors. As shown in Fig. 4, both NF-κB inhibitors, BAY11–7085 and parthenolide, caused similar concentration-dependent and complete inhibition of IL-8 release stimulated by GM-CSF or TNF-α . IC50 values for the inhibition of TNF-α - and GM-CSF-stimulated IL-8 release were 1.9 μ M and 1.8 μ M for BAY11–7085 and 1.7 μ M and 1.4 μ M for parthenolide, respectively (Fig. 4A, B). These data are in contrast to the partial inhibition of agonist-stimulated IL-8 release obtained with LY294002 and the p42/44MAPK inhibitor PD098059 (Fig. 4C, D). Moreover, when neutrophils were incubated concurrently with both cytokines, GM-CSF (10 ng/ml) and TNF-α  (200 U/ml), there was a synergistic increase in IL-8 release to values sevenfold higher than those achieved with either cytokine alone. LY294002 (1–10 μ M) and PD098059 (30–100 μ M) significantly attenuated IL-8 release stimulated by the combined cytokines when compared to match controls (Fig. 4E, F). The combined inhibition of PI3-kinase and p42/44MAPK pathways had no further inhibitory effect on cytokine-stimulated IL-8 release (data not shown).

2.5 Determining the relative contribution of cytokine-stimulated IL-8 release to GM-CSF- or TNF-α -induced neutrophil survival

In view of the above data and the reported capacity of IL-8 to inhibit neutrophil apoptosis 34, we hypothesized that the generation and autocrine action of IL-8 may contribute to the survival effect of GM-CSF or TNF-α . Using a specific neutralizing antibody, we confirmed a concentration-dependent inhibition of IL-8-mediated neutrophil survival at both 6 (p=0.02) and 20 h (p=0.003). In the presence of TNF-α , the antibody enhanced the early (6 h) killing effect of TNF-α  (TNF-α  18.4±0.5%, TNF-α /anti-IL-8 26.7±1.0%, p>0.0001; Fig. 5A) and inhibited the TNF-α  survival effect at 20 h (control 82.1±0.7%, TNF-α  64.9±4.4%, TNF-α / anti-IL-8 83.8±3.8%, p=0.013; Fig. 5B). In contrast, this strategy has no effect on GM-CSF-stimulated neutrophil survival at either 6 (% apoptosis, GM-CSF 1.3±0.2%, GM-CSF/anti-IL-8 2.6±0.8%, p=0.792) or 20 h (GM-CSF 26.7±3.1%, GM-CSF/anti-IL-8 30.2±3.1%, p=0.31) (Fig. 5A, B). Of note, the IL-8-neutralizing antibody alone had no affect on the basal rate of apoptosis at either 6 or 20 h.

To address concerns that the concentration of GM-CSF used might be supra-maximal and thereby mask any inhibitory effect of the IL-8-neutralizing antibody, the effect of this antibody was retested using a lower range of GM-CSF concentrations. Analysis at 20 h indicated that neutralization of IL-8 did not affect GM-CSF-stimulated survival even at growth factor concentrations as low as 0.1 ng/ml (Fig. 6A); in contrast, the concentration-dependent survival observed with TNF-α  was completely inhibited by the IL-8-neutralizing antibody at all TNF-α  concentrations (Fig. 6B).

Figure 4.

The synergistic roles of NF-κB, PI3-kinase and MAP kinase in GM-CSF- or TNF-α -stimulated IL-8 synthesis from human neutrophils. Neutrophils were incubated in the presence or absence of GM-CSF (10 ng/ml) or TNF-α  (200 U/ml) for 120 min. Supernatants were analyzed by ELISA for the synthesis and release of IL-8. (A, B) Concentration-dependent inhibition of IL-8 release from human neutrophils stimulated by either GM-CSF or TNF-α  by BAY11–7085 or parthenolide, respectively, (C) by the PI3-kinase inhibitor LY294002, or (D) by the MAP kinase inhibitor PD098059. Concentration-dependent inhibition of IL-8 release from neutrophils co-incubated with both GM-CSF (10 ng/ml) and TNF-α  (200 U/ml) by (E) LY294002 or (F) PD098059. Statistical analysis of controls via paired t-test (* or # p<0.05). Statistical analysis of concentration response curves via two-way ANOVA test (* or # p<0.05) compared to matching controls.

Figure 5.

Role of IL-8 in GM-CSF-stimulated neutrophil survival. Human neutrophils were incubated in the presence or absence of GM-CSF (10 ng/ml), TNF-α  (200 U/ml) or IL-8 (100 ng/ml) with or without a neutralizing antibody to IL-8 (concentration response curve 0.01–3 μ g/ml) for 6 (A) or 20 h (B). Cells were harvested and stained with May/ Grünwald/Giemsa and analyzed by light microscopy. Neutralizing anti-IL-8 antibody did not affect GM-CSF-stimulated neutrophil survival at 6 or 20 h. However, neutralization of IL-8, concentration-dependently enhanced TNF-α -stimulated apoptosis at 6 h and inhibited TNF-α -induced survival at 20 h. Data represent the percentage of apoptosis from two independent experiments, each performed in triplicate, together with mean ± SD values. Statistical analysis of controls by paired t-test (*, #, or † p<0.05). Statistical analysis of concentration response curves by two-way ANOVA test (# or † p<0.05).

Figure 6.

Neutralizing IL-8 inhibits TNF-α -dependent neutrophil survival. Human neutrophils were incubated with or without neutralizing antibody to IL-8 (3 μ g/ml) in the presence or absence of (A) GM CSF (concentration response curve 0.03–10 ng/ml) or (B) TNF-α  (concentration response curve 0.3–200 U/ml) for 20 h. Cells were harvested and assessed for apoptosis via FACS analysis. Neutralization of IL-8 did not affect GM-CSF-stimulated neutrophil survival at any concentration. In contrast, neutralizing anti-IL-8 antibody inhibited the TNF-α  concentration-dependent survival. Data shown as a single representative experiment.

3 Discussion

The potential synergistic effects and the interplay between the PI3-kinase and the NF-κB signaling pathways with respect to cell survival have proven to be extremely controversial. Much of this uncertainty relates to an excessive dependence on cell line and artificial expression systems 3739. This study has examined the role of NF-κB in GM-CSF-stimulated survival in freshly isolated primary human neutrophils, which we have previously shown to be PI3-kinase dependent 6. Our study shows that TNF-α  causes a robust activation of NF-κB via a PI3-kinase-independent activation of IκB degradation, leading to the subsequent NF-κB-dependent release of IL-8. This latter effect appears to significantly contribute to the ability of TNF-α  to modulate apoptotic thresholds in neutrophils 7. In contrast, GM-CSF causes a similar NF-κB-dependent stimulation of IL-8 release, but via a mechanism that does not involve IκB degradation or enhanced NF-κB activity, and this has little if any direct involvement in the pronounced survival effect of GM-CSF. Therefore, our data would suggest that the previously stated role of PI3-kinase signaling in TNF-α -stimulated NF-κB activation seen in 293 embryonic kidney cells 13, 37, 3T3 and endothelial cells 38, 39, may have been overestimated. However, a recent paper has described GM-CSF-stimulated NF-κB activation in adherent but not suspended neutrophils, suggesting the potential for PI3-kinase and NF-κB cross-talk in these cells, although this is not operational in circulating cell 35.

Several reports in cell lines and overexpression systems have postulated important links between the PI3-kinase and NF-κB signaling pathways. The exact point of interaction and the precise relevance of these overexpression systems to the events occurring in primary cells remain controversial. Ozes and colleagues 13 have shown that TNF-α -stimulated NF-κB activation in HEK293 cells requires the parallel stimulation of NF-κB-inducing kinase (NIK) and PI3-kinase/Akt. The later PI3-kinase step was proposed to activate IκK through phosphorylation of IκKα  at threonine 23 (T23) and suggested to be essential for IκK and NF-κB activity. Moreover, TNF-α  has been reported to induce PI3-kinase activity in p85 immunoprecipitates leading to phosphorylation of Akt 40, and wortmannin (a PI3-kinase inhibitor) was shown to completely inhibit TNF-α -stimulated IκB degradation. This hypothesis proposed that some of the anti-apoptotic activity of Akt is directed through NF-κB signaling, which is in itself a major regulator of apoptotic thresholds. However, this mechanism does not appear to pertain to the neutrophil. We have previously shown that while TNF-α  priming enhances fMLP receptor-stimulated PtdIns(3,4,5)P3 accumulation, TNF-α  alone has no effect on PtdIns(3,4,5)P3 accumulation 8, and we now show that LY294002 has no effect on TNF-α -stimulated IκB loss. In contrast, GM-CSF, which activates PI3-kinase and elevates PtdIns(3,4,5)P3 in human neutrophils, was unable to initiate IκB degradation. Together, these data suggest that in the neutrophil, TNF-α  induces IκB degradation and NF-κB activation via a PI3-kinase-independent mechanism and that the activation of PI3-kinase by GM-CSF is insufficient to induce IκB degradation.

In these studies, we have used two structurally independent, site-specific inhibitors of NF-κB to distinguish potential sites of interaction between PI3-kinase and NF-κB pathways. BAY11–7085 inhibits TNF-α -stimulated IκB degradation, while parthenolide inhibits phosphorylation of the p65 transactivation domain, thereby preventing the formation of the NF-κB/DNA complex 41. In agreement with our previous observations, BAY11–7085 and parthenolide, when used at moderately high concentrations (>10 μ M), increased the extent of basal apoptosis; this is consistent with the demonstration that the NF-κB pathway is constitutively activated in these primary cells and that this activity is important in suppressing the constitutive rate of apoptosis 42. At submaximal concentrations, BAY11–7085 and parthenolide display a differential effect on cytokine-mediated survival by enhancing the early pro-apoptotic effect of TNF-α  and abrogating its late survival effect while having no effect on the GM-CSF stimulated survival. Hence, GM-CSF appears to enhance neutrophil survival by a route not involving NF-κB activation. In contrast, NF-κB activation appears to be fundamentally important in TNF-α -stimulated survival and in determining the intensity of the early pro-apoptotic effect. These data would again concur with our previous studies 5 showing that the inactivation of NF-κB by gliotoxin enhanced TNF-α -stimulated neutrophil apoptosis by a caspase-dependent effector pathway. Pre-incubation of human neutrophils with the caspase inhibitor zVAD-fmk (10 μ M) completely inhibited TNF-α - and gliotoxin-stimulated cell death. Moreover, cycloheximide initiated a similar response to that observed with gliotoxin, suggesting that NF-κB activity delays neutrophil cell death through the production of a newly synthesized survival factor rather than through any direct effect on apoptotic effector mechanisms.

One potential result of NF-κB/PI3-kinase interaction that may stimulate the production of a neutrophil-derived survival factor is the synthesis and release of IL-8. We have confirmed that TNF-α  and GM-CSF are both capable of stimulating IL-8 release from neutrophils 34, 36 and revealed a major synergistic increase in IL-8 production in the presence of both agents. The signaling basis of the IL-8 response to GM-CSF is less certain. While GM-CSF does not directly stimulate IκB degradation, both BAY11–7085 and parthenolide cause a concentration-dependent and complete inhibition of GM-CSF-stimulated IL-8 synthesis from these cells, a finding that confirms a central role for NF-κB in IL-8 transcriptional regulation. However, in the case of GM-CSF, this clearly occurs in the absence of enhanced IκB degradation. The ability of GM-CSF to stimulate IL-8 synthesis by an IκB-independent pathway may be explained by the additional input of PI3-kinase and ERK1/2 pathways to basal IL-8 production (see Fig. 4C–F). This view is supported by the ability of LY294002 and PD098059 to partially inhibit GM-CSF-stimulated IL-8 synthesis and by the synergistic effect of GM-CSF and TNF-α  in stimulating IL-8 synthesis. ERK1/2 phosphorylation has been shown to increase the affinity of AP-1 binding to its consensus sequence in promoter regions 43, 44, and the IL-8 promoter is known to contain AP-1-binding domains, which function as basal level enhancers increasing IL-8 transcription following NF-κB activation 43. The latter study confirmed that inhibition of ERK1/2 caused only partial inhibition of IL-8 synthesis in GM-CSF-stimulated cells. Hence, maximal NF-κB transactivation involves not only its translocation to the nucleus and assembly of the transcription complex, but additional phosphorylation inputs from PI3-kinase and MAP kinase. Of note, although GM-CSF stimulates the synthesis and release of significant amounts of IL-8, the predicted contribution of any autocrine survival effect of IL-8 towards GM-CSF-stimulated (PI3-kinase-dependent) survival would be marginal as IL-8 has a far more modest survival effect.

Further analysis of the contribution made by the autocrine effect of IL-8 towards GM-CSF- and TNF-α -stimulated survival was investigated using an IL-8-neutralizing antibody. This identified a striking difference between GM-CSF and TNF-α ; with the IL-8 antibody having no effect on GM-CSF-stimulated survival at 6 or 20 h, but causing significant enhancement of TNF-α -stimulated apoptosis at 6 h and complete reversal of its survival effect at 20 h 34. These data again support the view that NF-κB stimulation is not a central pathway underlying the GM-CSF survival effect.

In summary, the present study has investigated the dual roles of the PI3-kinase and NF-κB signaling pathways in GM-CSF- and TNF-α -stimulated human neutrophil survival. As illustrated in Fig. 7, our data indicate that TNF-α -induced survival in human neutrophils depends on the autocrine effects of IL-8 generated as a direct response to IκB/NF-κB activation. In contrast, GM-CSF stimulation causes a more profound survival effect that is directly PI3-kinase dependent and independent of its capacity to stimulate IL-8 release. The mechanism whereby GM-CSF stimulates NF-κB-dependent IL-8 synthesis in the absence of IκB degradation remains to be determined.

Figure 7.

 Schematic diagram depicting the potential points of cross-talk between the PI3-kinase and NF-κB signaling pathways in GM-CSF- or TNF-α -stimulated neutrophil survival.

4 Materials and methods

4.1 Neutrophil preparation and culture conditions

Human neutrophils were purified from the peripheral blood of healthy non-atopic volunteers as described 45. Freshly isolated neutrophils were suspended at 5×106 cells/ml in Iscove's modified Dulbecco's medium (MDM) (Invitrogen, Paisley, GB), supplemented with 10% autologous serum and 100 U/ml penicillin and 100 μ g/ml streptomycin (Sigma, GB), and cultured in 96-well ultra-low attachment plates (Costar, Bucks, UK) in a humidified 5% CO2 atmosphere at 37°C. To examine the effect of various cytokines on the rate of constitutive neutrophil apoptosis, cells were incubated in the presence or absence of predetermined optimal concentrations of GM-CSF (10 ng/ml), TNF-α  (200 U/ml) or IL-8 (100 ng/ml) (R & D, Abingdon, GB) for the indicated time periods (6 and 20 h). These time periods were previously shown to reveal the early pro-apoptotic effect of TNF-α  and the later survival effect of both GM-CSF and TNF-α  in cells that remain fully viable (i.e. trypan blue excluding) over this time course 23, 26. Where neutralizing antibodies or selective pathway inhibitors were used, neutrophils were pre-incubated with LY294002 (0.03–10 μ M), PD98059 (0.03–100 μ M), parthenolide (0.1–20 μ M), BAY11–7085 (0.1–20 μ M) (Merck Biosciences Ltd., GB); IL-8-neutralizing antibody (0.01–30 μ g/ml) (R & D) or appropriate vehicle for 20 min at 37°C prior to the addition of the above cytokines.

4.2 Morphological analysis of neutrophil apoptosis

Cytocentrifuged cells were fixed and stained as previously described 6. Morphological analysis of neutrophil apoptosis was assessed under oil immersion light microscopy (×100 objective) with the observer blinded to the experimental conditions. Apoptotic neutrophils were clearly identifiable and defined as cells with darkly stained condensed nuclei. For each of the conditions investigated, slides were prepared from triplicate wells and a minimum of 400 neutrophils counted per slide.

4.3 Annexin V analysis of neutrophil apoptosis

Neutrophils were cultured as detailed then centrifuged (275 × g, 5 min at 4°C), and the cell pellet was resuspended in 200 μ l Hepes buffer (10 mM Hepes-NaOH pH 7.4, 150 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2) containing Annexin V-FITC (1 μ g/ml) and propidium iodide (10 μ g/ml). The samples were incubated for 30 min at 4°C in the dark and the volume was increased to 500 μ l with Hepes buffer immediately before analysis by flow cytometry (FACSCalibur, Becton Dickinson, Oxford, GB).

4.4 SDS-PAGE and immunoblotting

Neutrophils were incubated for varying times with or without GM-CSF (10 ng/ml) or TNF-α  (200 U/ml) in the presence or absence of inhibitors (as detailed in the figure legends). The cells were then pelleted and resuspended in lysis buffer [10 mM Tris-HCl pH 7.8, 1.5 mM EDTA, 10 mM KCl, 0.5 mM DTT, 1 mM sodium orthovanadate, 2 mM levamisole, 0.5 mM benzamidine, 0.05% NP-40 and proteinase inhibitor cocktail (Roche, East Sussex, GB)] and incubated on ice for 10 min prior to and following sonication (Soniprep 150, MSE, setting 13 for 15 s). Cellular debris was pelleted by centrifugation at 10,000 × g for 15 min at 4°C. Supernatants were separated on 12% (w/v) polyacrylamide gels and separated proteins were electrophoretically transferred to PVDF membranes. The membranes were blocked, then incubated overnight at 4°C in PBS-Tween-20 with polyclonal antibody to IκBα  (New England Biolabs, Hertfordshire, GB) at dilutions of 1:1,000. A peroxidase-conjugated secondary antibody (final dilution 1:2,000) followed by detection with enhanced chemiluminescence (Pierce, GB) and subsequent exposure to X-ray films (XOMAT-AR, Kodak, GB) provided a sensitive readout.

4.5 NF-κB activity assay

Neutrophils were incubated with or without GM-CSF (10 ng/ml) or TNF-α  (200 U/ml) for 2 h. Cells were then pelleted and resuspended in NucBuster protein extraction lysis buffer (CN Biosciences, Nottingham, GB), to extract the nuclear fraction according to the manufacturer's instructions. The nuclear fraction was analyzed using a TransAM NF-κB family transcription factor assay kit according to the manufacturer's instructions (ActiveMotif, Belgium). Briefly, in a 96-well plate the active form of NF-κB in the extract binds to the oligonucleotide containing the NF-κB consensus site (GGGACTTTCC). Primary antibody detection of p50 and p65 subunits was achieved with a HRP-conjugated secondary antibody, providing a colorimetric readout.

4.6 IL-8 ELISA

Cell suspensions were aspirated, centrifuged for 5 min at 2,000 × g (4°C), and the supernatants were stored at –80°C until further analysis. IL-8 was measured by an ELISA developed in house: 96-well flat-bottom high binding micro-ELISA plates (Greiner, GB) were coated with 2 μ g/ml mouse monoclonal anti-human IL-8 antibody (R & D) in carbonate-bicarbonate buffer pH 9.6 for 2 h at room temperature, using a working volume of 50 μ l. Plates were washed three times with PBS containing 0.05% Tween-20 (PBS-T) and blocked with 5% heat-inactivated FCS (Sigma) in PBS-T for 1 h at room temperature. After one wash with PBS-T, aliquots of standards (recombinant human IL-8; R & D) and samples (supernatants) with the relevant controls were added and incubated overnight at 4°C. Following three washes with PBS-T, biotinylated goat anti-human IL-8 antibody diluted to 0.25 μ g/ml in PBS-T + 5% FCS was added and incubated for 2 h at room temperature. Plates were washed three times and ExtraAvidin® alkaline phosphatase conjugate (1:400) (Sigma) in PBS-T containing 5% FCS was added and incubated for a further 2 h at room temperature. After three final washes, p-nitrophenylphosphate (Sigma) at 1 mg/ml in diethanolamine buffer pH 9.8 was added, and the plates were read at 405 nm in an automated plate reader (3550, Bio-Rad, GB). Results were analyzed using Microplate Manager software (Bio-Rad). The lower level sensitivity of this assay was 10 pg/ml.

4.7 RT-PCR procedures

Total RNA was isolated using RNeasy mini kits (Qiagen, West Sussex, GB). RNA (2 μ g) was transcribed into cDNA using oligo(dT) primers (Invitrogen) and 50 U of reverse transcriptase (Promega, GB). PCR amplification was performed using specific primer sets for IκBα  (sense: 5′-ACC-TCC-ACT-CCA-TCC-TGA-AG; anti-sense: 3′-CAC-ACA-GTC-ATC-ATA-GGG-CAG; 376-bp product), for control reactions a specific primer set for β -actin (sense: 5′-GTG-GGG-CGC-CCC-AGG-CAC-CA; anti-sense: 3′-CTC-CTT-AAT-GTC-ACG-CAG-CAC-GAT-TTC; 548-bp product) was used. PCR (35 cycles) was performed using 2 U of ampliTaq DNA polymerase (Bioline, GB). PCR products were analyzed by agarose gel electrophoresis and imaged with ethidium bromide under UV light.

4.8 Statistical analysis

Data are presented as mean ± SEM for each experimental group unless otherwise stated. Two-way analysis of variance was employed for concentration-dependent curves, or Student's t-test for comparison between two groups. Values of p<0.05 were considered significant.

Acknowledgements

This work was financially supported by The Wellcome Trust, British Lung Foundation and the Medical Research Council.

Footnotes

  1. 1

    WILEY-VCH

  2. 2

    WILEY-VCH

  3. 3

    WILEY-VCH

  4. 4

    WILEY-VCH

  5. 5

    WILEY-VCH

  6. 6

    WILEY-VCH

  7. 7

    WILEY-VCH

Ancillary