Sodium Salicylate Promotes Neutrophil Apoptosis by Stimulating Caspase-Dependent Turnover of Mcl-1¹

Apoptosis of effete immune cells is essential for the successful resolution of inflammation. Neutrophils are unique among immune cells in that they have extremely high rates of constitutive apoptosis, and have a half-life estimated to be only 12–18 h (1, 2, 3, 4, 5). During inflammation, exposure of neutrophils to proinflammatory signals such as cytokines, can delay, but not halt this apoptosis so that neutrophils at sites of inflammation will have an extended survival that is beneficial for them to perform their immune functions. Apart from proinflammatory cytokines such as GM-CSF, IFN-γ, IL-1, a variety of other factors such as LPS, integrin engagement, and hypoxia, have all been reported to delay apoptosis in disease (1, 2, 6, 7, 8, 9). However, in some inflammatory diseases, such as inflammatory lung disease, dysregulated neutrophil apoptosis is implicated in enhanced neutrophil survival and neutrophil-mediated tissue damage (10, 11). Understanding the mechanisms that regulate neutrophil apoptosis and extended survival could lead to new insights into disease processes and hence new possibilities for therapeutic intervention in human diseases in which impaired neutrophil apoptosis underlies disease pathology.

Much research now implicates Mcl-1 as a key regulator of neutrophil apoptosis and survival (4, 6, 12, 13). Mcl-1 is a relatively understudied antiapoptotic member of the Bcl-2 family that has some unusual properties compared with other family members. Mcl-1 was first isolated as an early-induction gene during myeloid cell differentiation, and many studies have shown it to enhance cell survival in transfection studies (14, 15, 16). Mcl-1 is a much larger protein than Bcl-2 and Bcl-x_L, partly because it contains PEST sequences and other motifs that target it for proteolysis by the proteasome (14, 17). Indeed, the half-life of Mcl-1 is only 2–3 h, depending on the cell type and the culture conditions (12, 15, 17). Although Mcl-1 levels can be increased in cells by activated transcription and translation, profound changes in its cellular levels can occur by changes in its rate of turnover (18, 19). For example, inhibition of the proteasome can block Mcl-1 turnover and extend neutrophil survival (19), whereas differential posttranslational modifications of the protein by phosphorylation can either stabilize the protein or else further accelerate its normally high rate of turnover. For example, phosphorylation of Thr¹⁶³ by Erk can decrease the turnover of Mcl-1, while hyperphosphorylation of the protein by Taxol or okadaic acid can further enhance turnover (19, 20). These changes in Mcl-1 turnover are reflected by changes in the rate of cellular apoptosis, indicating that they are of pathological importance in determining cell fate in disease. More recently, however, caspase-dependent proteolysis of Mcl-1 has been reported in a number of cell types (21, 22, 23, 24), but in neutrophils undergoing constitutive or cytokine-delayed apoptosis, there is no evidence for caspase-dependent proteolysis of Mcl-1 (19).

In the present study, we have examined the effects of sodium salicylate on neutrophil apoptosis and Mcl-1 expression. Sodium salicylate, the nonacetylated derivative of aspirin, is a poor inhibitor of cyclooxygenase but is a potent inducer of apoptosis in a number of cell types (25, 26, 27). Its mode of action is undefined, but it has been reported to affect a variety of intracellular signaling systems including NF-κB and a number of kinase pathways, including the MAPKs (28, 29, 30, 31, 32, 33). It has also been reported to result in decreased levels of Mcl-1 protein in TF-1 cells, assumed to be via effects on mcl-1 transcription and mRNA stability (27). In this study, we show that sodium salicylate is a potent inducer of apoptosis of human neutrophils and completely blocks the protective effects that the cytokine GM-CSF normally has on delaying neutrophil apoptosis. Sodium salicylate decreased mcl-1 (but not bfl-1) transcription and enhanced the rate of turnover of Mcl-1 protein. In contrast to constitutive or GM-CSF-regulated Mcl-1 turnover that was mediated via the proteasome, sodium salicylate-induced turnover was largely caspase dependent. Sodium salicylate also induced profound changes in the activities of a number of kinases implicated in neutrophil survival, especially a marked and sustained activation of p38-MAPK. These data thus present novel insights into the mechanisms by which sodium salicylate can accelerate neutrophil apoptosis that may have implications for the treatment of inflammatory diseases.

RPMI 1640 medium and FCS were from Invitrogen Life Technologies. Anti-human Mcl-1 mAb (13656E), RiboQuant MultiProbe RNase Protection System, and hAPO-2c Template set were from BD Pharmingen. MG-132, PD98089, SB202190, wortmannin, ALLN (N-acetyl-leucinyl-leucinyl-norleucinal), MDL 28170 (carbobenzoxy-valinyl-phenylalaninal), lactacystin, Z-VAD-FMK, okadaic acid, LY294002, and phosphatase inhibitor set II (imidazole, sodium fluoride, sodium molybdate, sodium orthovanadate, sodium tartrate dihydrate) were from Calbiochem. Cycloheximide, actinomycin D, pooled human AB serum, and sheep anti-mouse IgG were from Sigma-Aldrich. GM-CSF was from Roche Applied Science. FITC-Annexin V was from BioSource International. Anti-phospho p44/42 Thr²⁰²/Tyr²⁰⁴ (ERK1/2), anti-phospho p38-MAPK Thr¹⁸⁰/Tyr¹⁸², anti-phospho Akt Ser⁴⁶³, and anti-caspase-3, -8, -9, and -10 Abs were from Cell Signaling Technology. Donkey anti-rabbit IgG and ECL detection kit was from Amersham Biosciences. Donkey anti-goat IgG and anti-Mcl-1 (S-19) were from Santa Cruz Biotechnology. Polymorphoprep was from Robbins Scientific.

Neutrophils were isolated from heparinized venous blood from healthy volunteers by one-step centrifugation through Polymorphprep (34, 35). After hypotonic lysis to remove contaminating erythrocytes, cells were resuspended in RPMI 1640 medium supplemented with 5% pooled human male AB serum at 5 × 10⁶/ml. Culture was at 37°C in polypropylene tubes with gentle agitation. Purity and viability were routinely >95%, as assessed by May-Grünwald-Giemsa staining and trypan blue exclusion, respectively. GM-CSF was added at 50 ng/ml. The following additions were also made: cycloheximide, 10 μg/ml; MG-132, 50 μM; okadaic acid, 1 μM; PD98059, 50 μM; SB202190, 1 μM; wortmannin, 20 nM; H-89, 5 μM; LY294002, 10 μM; ALLN, 100 μM; carbobenzoxy-valinyl-phenylalaninal (MDL 2810), 1 μM; Z-VAD-FMK, 50 μM; lactacystin, 40 μM.

Following culture, a 20-μl aliquot of suspension was made up to 200 μl with RPMI 1640, and cells were cytocentrifuged using a Shandon Cytospin3 (Runcorn). Romanowsky staining of cytospins allowed apoptosis to be scored by morphology, as described in Ref.36 . This method correlates well with other markers of apoptosis (37).

Following culture, 10⁶ cells were rapidly lysed in boiling reducing SDS-PAGE sample buffer containing aprotinin (20 μg/ml), leupeptin (20 μg/ml), pepstatin (10 μg/ml), PMSF (400 μg/ml), and phosphatase inhibitor set II (1/100 dilution), immediately boiled for 5 min with occasional vortexing, and stored at −80°C until use. SDS-PAGE and electrotransfer to polyvinylidene difluoride membranes were performed as described (12). Primary Abs used were: Mcl-1 (13656E), anti-phospho p44/42 (ERK), anti-phospho p38-MAPK, anti-phospho Akt, anti-Bax. HRP-conjugated secondary antisera used were the following: donkey anti-rabbit IgG and donkey anti-goat IgG. Bound Abs were detected using the ECL system. Densitometry on carefully exposed blots (to avoid film saturation) was performed with Image 1.44 VDM software (National Institutes of Health). Ponceau S-stained actin on membranes after electrotransfer was measured to confirm equivalence of loading of neutrophil samples.

Following culture of cells, total RNA was extracted using TRIzol reagent (Invitrogen Life Technologies) following the manufacturers’ instructions. Typical RNA yields were ∼4 μg/10⁷ neutrophils. For RPA, RNA from 15 × 10⁶ neutrophils was used per sample. RiboQuant MultiProbe RNase Protection System and hAPO-2c Template set were used for all RPAs using [³³P]UTP for probe labeling. The manufacturers’ instructions were followed for all steps of the RPA. Protected fragments were separated on 30-cm 5% acrylamide gels, and radioactivity in fixed, dried gels was detected with a Bio-Rad GS-363 Molecular Imager. Protected fragments were identified by comparison to native probes and HeLa RNA supplied with the RPA kit. All bands were quantified using Molecular Analyst software supplied with the molecular imager. Quantified data were expressed relative to signals obtained from the housekeeping gene L32.

Data sets were analyzed using the paired Student’s t test.

Previous work has shown that sodium salicylate can induce apoptosis in a variety of myeloid cells, including neutrophils (25, 26, 27). Because GM-CSF can promote neutrophil survival, it was first important to determine whether sodium salicylate had any effects on GM-CSF-induced neutrophil survival, because this would indicate whether salicylate could affect the apoptosis kinetics of inflammatory neutrophils. Neutrophils were therefore incubated in the presence and absence of GM-CSF with increasing concentrations of sodium salicylate. Apoptosis was determined by cell morphology at 6 and 22 h, and Mcl-1 protein levels were determined by Western blotting after 8-h incubation.

As has previously been reported, GM-CSF significantly delayed neutrophil apoptosis (Fig. 1,A). Over 80% of the neutrophils showed apoptotic morphology after 22-h incubation in the absence of this cytokine, whereas only 30% were apoptotic in the presence of GM-CSF. The addition of 10 mM sodium salicylate significantly accelerated the rate of constitutive neutrophil apoptosis: by 6-h incubation, 35% (±3%) of the cells were apoptotic in the presence of sodium salicylate, compared with only 15% (±3%) of the cells in the absence of this agent (p < 0.05). Remarkably, the effects of sodium salicylate on apoptosis were much greater when the cells were incubated in the presence of GM-CSF. For example, by 6-h incubation, apoptosis in the presence of GM-CSF increased from 10% (±2%) to 23% (±1%) in 1 mM, and to 50% (±6%) in 10 mM sodium salicylate (p < 0.05). These enhanced rates of apoptosis in the presence of sodium salicylate were accompanied by dramatic decreases in the cellular levels of Mcl-1, a key antiapoptotic protein regulating neutrophil survival. Levels of Mcl-1 were significantly decreased in the cells incubated in the presence of 10 mM sodium salicylate. Representative Western blots are shown in Fig. 1,B and indicate that, even by 2-h incubation, Mcl-1 protein levels were slightly decreased by addition of sodium salicylate, and by 5-h incubation with this agent were barely detectable. Decreases in Mcl-1 protein levels therefore preceded the appearance of apoptotic morphology (Fig. 1 B).

The above observations indicate that sodium salicylate can exert dramatic effects on the cellular levels of Mcl-1, a protein with a very high turnover rate. The effects of this drug on mRNA levels were therefore measured using an RPA that simultaneously measured mRNA levels of a number of other antiapoptotic family members. Control (untreated) neutrophils expressed high levels of mRNA for Mcl-1 and Bfl-1 (A1), another antiapoptotic member of the Bcl-2 family (Fig. 2 A). Transcripts for Bad, Bak, Bik, Bax, and Bcl-2 were extremely low or negligible. GM-CSF treatment resulted in increases in the Mcl-1 and Bfl-1 signals that were 30% (±10%) and 85% (±11%), respectively, above control, untreated values (p < 0.05). Sodium salicylate had no effect on either the control or GM-CSF-regulated signal for Bfl-1, but significantly decreased Mcl-1 mRNA levels, both in control and cytokine-treated cells.

This observation may indicate that sodium salicylate selectively blocks Mcl-1, but not Bfl-1 transcription. However, a possibility that must be ruled out is that sodium salicylate blocks transcription of both genes (and perhaps other genes), but the relative levels of transcripts detected after 3-h incubation reflected their relative turnover rates. For example, transcripts with low turnover rates would be detected at higher levels than those transcripts with a higher rate of turnover after the termination of transcription.

The stabilities of Bfl-1 and Mcl-1 mRNA were therefore measured in experiments in which transcription was blocked by use of actinomycin D and transcript levels detected 1, 3, and 5 h after transcriptional block. For comparison, the stability of mRNA for Bcl-x_L, which is constitutively expressed in neutrophils (12), was also measured. In neutrophils blocked in transcription, the levels of mRNA for Bcl-x_L (measured by the RPA) changed very little over the 5-h incubation period, indicating that, in these circumstances, this transcript is extremely stable and has a relatively long half-life (Fig. 2 B). In contrast, mRNA levels for both Bfl-1 and Mcl-1 declined rapidly in neutrophils blocked in transcription. Thus, these two transcripts have similar, very short half-lives, estimated to be ∼2 h.

Mcl-1 has a rapid turnover rate (14, 17), because it contains PEST sequences and other motifs that target it for proteolysis. Its half-life is reported to be ∼2–3 h, depending on the cell type and culture conditions, but in neutrophils this half-life can be either enhanced (e.g., by cytokine signaling) or decreased further (e.g., by okadaic acid treatment that results in its hyperphosphorylation). Neutrophils were therefore incubated with the protein translation inhibitor cycloheximide to block de novo biosynthesis, and Mcl-1 levels remaining in cells were measured by Western blotting. Fig. 3 shows that, in cycloheximide-treated cells, Mcl-1 protein levels declined rapidly, so that, by 3 h, only ∼50% of the original levels remained. However, in the presence of sodium salicylate, Mcl-1 turnover was significantly increased, with its half-life estimated to be 90 min.

Previous work has shown that GM-CSF treatment can increase the stability of Mcl-1, thereby increasing its cellular levels by decreasing its rate of turnover (19). Fig. 4 shows that sodium salicylate increased the rate of degradation of both control and GM-CSF-treated neutrophils. Thus, the stability increase normally induced by GM-CSF was abrogated by the action of sodium salicylate.

We have previously shown that inhibitors of the proteasome such as MG-132, lactacystin, and ALLN can all prevent Mcl-1 turnover in neutrophils and delay apoptosis (19). Recent reports have implicated caspase-dependent cleavage of Mcl-1 that may also contribute to its degradation in other cells under certain circumstances, and caspase-cleavage sites are identified by sequence analysis (21, 22, 23, 24). We therefore tested the effects of inhibitors of caspases, calpains and the proteasome, for their ability to regulate constitutive and salicylate-enhanced Mcl-1 turnover.

Fig. 5,A shows that the proteasome inhibitor MG-132 (or lactacystin and ALLN; data not shown) blocked Mcl-1 turnover in cycloheximide-treated cells. However, the pancaspase inhibitor Z-VAD-FMK (and calpain inhibitors; data not shown) had no effect on this rate of turnover. However, in the presence of sodium salicylate (Fig. 5,B), the proteasome inhibitor MG-132 offered some protection against turnover 2 h after the addition of cycloheximide, but by 5 h it was completely ineffective at preventing salicylate-induced turnover. In contrast, the pancaspase inhibitor Z-VAD-FMK significantly delayed salicylate-induced turnover of Mcl-1 at both 2 and 5 h. Representative blots are shown in Fig. 5 C. Z-VAD-FMK also protected neutrophils from salicylate-induced apoptosis. For example, after 5-h incubation, sodium salicylate treatment resulted in 31% (±4%; n = 5) of the neutrophils exhibiting apoptotic morphology, compared with control (untreated) cells (12 ± 3%; n = 5). In the presence of Z-VAD-FMK plus sodium salicylate, apoptosis was only 8% (±2%; n = 5; p < 0.01). Even more dramatic effects of Z-VAD-FMK were observed in cycloheximide-treated cells with rates of apoptosis being the following (n = 5): cycloheximide alone, 39% (±9%); cycloheximide plus sodium salicylate, 85% (±3%); cycloheximide plus sodium salicylate plus Z-VAD-FMK, 25% (±25%). Thus, the accelerated apoptosis induced by sodium salicylate was caspase dependent.

These data indicate that, in control cells, the proteasome is the major route for Mcl-1 turnover, whereas salicylate induces an alternative caspase-dependent turnover of Mcl-1.

The above experiments indicated that sodium salicylate triggered caspase-mediated turnover of Mcl-1 because of the ability of the pancaspase inhibitor to prevent turnover. It was therefore necessary to determine 1) whether caspase-cleaved products of Mcl-1 could be detected following salicylate treatment and 2) which caspases were activated by sodium salicylate. Fig. 6 A shows Western blots of neutrophils treated with sodium salicylate and also HeLa cells treated with staurosporine and sodium salicylate to induce apoptosis. A band at 28 kDa was detected in HeLa cells treated for 24 h with staurosporine, whereas bands at 28 and 24 kDa were detected following sodium salicylate treatment. These are attributable to caspase-cleaved products and were not detected in cells treated with the pancaspase inhibitor (data not shown). In neutrophils incubated for 3 h with sodium salicylate, caspase-cleaved products were detected at ∼28 kDa, but levels of this cleavage product thereafter decreased. Caspase-cleaved products of Mcl-1 were not detected in cells treated with the pancaspase inhibitor (data not shown). Thus, in contrast to HeLa cells, the caspase-cleaved forms of Mcl-1 in neutrophils are also extremely unstable. Note that Ab S-19 was used to detect these caspase-cleaved forms of Mcl-1.

Fig. 6 B shows that caspase-3, -8, -9, and -10 were all activated within 2 h of addition of this agent. This is indicated by the decreased signal for procaspase-10, -9, and -3 (at 64, 47, and 32 kDa, respectively) and the decreased signal for full-length caspase-8 (at 57 kDa) and appearance of the cleaved forms at 41/43 kDa. Caspase-3 and -8 were also rapidly activated in control neutrophils, but GM-CSF treatment prevented this activation.

GM-CSF-mediated increases in stability of Mcl-1 are signaled via PI-3K/Akt and Mek/Erk pathways (19). In contrast, hyperphosphorylation of Mcl-1 can accelerate Mcl-1 turnover. We therefore tested the effects of sodium salicylate on activation levels of these signaling pathways in control cells and GM-CSF-treated cells.

Control (untreated) neutrophils exhibited only low or negligible levels of activation of Erk or Akt, as shown in Fig. 7, A and B. However, GM-CSF induced rapid, but transient increases in the activation (phosphorylation) status of these signaling systems, which peak by 5–10 min and return to basal levels by 4–5 h. Previous work (19) has shown that activation of these signaling pathways is crucial for GM-CSF-stimulated apoptosis delay and increased Mcl-1 stability. The addition of sodium salicylate had no effect on basal (unstimulated) levels of these two signaling pathways but markedly decreased activation induced by GM-CSF. Thus, sodium salicylate completely prevented GM-CSF-induced activation of Akt, and completely prevented activation of the higher molecular mass isoforms of Erk (but not the lower molecular mass isoform). The observations may explain the mechanism by which sodium salicylate blocks the effects of GM-CSF on Mcl-1 stability and apoptosis by blocking activation of these signaling systems.

Endogenous levels of activation of p38-MAPK are low in resting neutrophils, and are constitutively activated by 4- to 5-h incubation as the cells begin to enter apoptosis (Fig. 7 C). As has been reported previously, GM-CSF stimulates complex changes in activation of p38-MAPK (19). The addition of sodium salicylate to control neutrophils resulted in a very rapid and sustained activation of p38-MAPK that remained elevated throughout the time course of the experiment. The addition of GM-CSF had little additional effect on the very high levels of p38-MAPK activation observed after sodium salicylate addition.

The above experiments showed that sodium salicylate blocked GM-CSF activation of Erk and Akt, but stimulated p38-MAPK activation. These data could indicate cross talk in the regulation of activation of these signaling pathways. For example, these results could indicate that p38-MAPK activity may down-regulate Erk and Akt activation or, alternatively, that Akt and/or Erk may down-regulate p38-MAPK activity. The addition of SB202190 (to inhibit p38-MAPK) had no effect on either the endogenous or GM-CSF-regulated activation of Akt or Erk (data not shown). Similarly, the addition of PD98059 (to inhibit Mek/Erk) had little or no effect on endogenous or GM-CSF-regulated Akt activation (data not shown). However, PD98059 altered the activation kinetics of GM-CSF-stimulated p38-MAPK activation (Fig. 8,A). Furthermore, whereas LY294002 (a PI3K inhibitor) greatly enhanced and sustained GM-CSF-regulated p38-MAPK activation (Fig. 8 A), it had no effect on endogenous p38-MAPK activation (data not shown). Thus, inhibition of Akt partly mimicked the effects of sodium salicylate, particularly in the presence of GM-CSF, suggesting that PI3K/Akt down-regulates p38-MAPK via signaling cross talk.

We have previously shown that the PP2A phosphatase inhibitor, okadaic acid, results in accelerated Mcl-1 turnover and acceleration of neutrophil apoptosis (19), effects resembling those of sodium salicylate on neutrophils. We therefore tested the effects of okadaic acid on the activation kinetics of p38-MAPK. Fig. 8,B shows that okadaic acid treatment of neutrophils results in extensive and sustained activation of p38-MAPK in the presence and absence of GM-CSF, an effect remarkably similar to the effect of sodium salicylate (Fig. 7 C).

The above data indicate that the acceleration of neutrophil apoptosis stimulated by sodium salicylate, may be mediated via increased activation of p38-MAPK. Therefore, inhibiting the activity of p38-MAPK would be predicted to block the apoptosis-inducing effect of sodium salicylate. Neutrophils were incubated in the presence and absence of sodium salicylate or SB202190. However, this latter inhibitor had no effect on either constitutive or sodium salicylate-accelerated apoptosis (Fig. 9). Apoptosis measured by morphology after 8-h incubation of control (untreated) cells was 23% (±9%; n = 4), whereas apoptosis in the presence of SB202190 was 31% (±14%; n = 4). Similarly, SB202190 had no effect on salicylate-induced apoptosis measured at 8-h incubation (salicylate only, 79 ± 11%; salicylate plus SB202190, 82 ± 11%; n = 4). Okadaic acid treatment also resulted in extensive apoptosis in these experiments.

In this report, we show, for the first time, that sodium salicylate, a well-known inducer of apoptosis in a variety of cell types, greatly accelerates neutrophil apoptosis by increasing the rate of turnover of Mcl-1, a key survival protein in these cells. In view of the growing awareness that the antiapoptotic phenotype of many malignant cells is due to overexpression of Mcl-1, it is also likely that accelerated Mcl-1 turnover may play a fundamental role in the induction of apoptosis seen upon salicylate treatment of such cells. We also show that, although the proteasome is the major route for Mcl-1 turnover during constitutive or GM-CSF-regulated apoptosis, sodium salicylate treatment results in a caspase-dependent turnover of Mcl-1. We also show salicylate-induced changes in activities of several MAPKs, especially blocking of GM-CSF-induced activation of Erk and Akt, and increased and sustained activation of p38-MAPK, events that are likely to alter Mcl-1 turnover and affect apoptosis.

Mcl-1 is unusual among Bcl-2 family proteins in that it has a very short half-life, which is due to the fact that it contains PEST regions, Arg:Arg motifs, and other domains that target it for proteolysis (14). The proteasome is one route by which Mcl-1 may undergo degradation, and there have been several reports that show that proteasome inhibitors can block Mcl-1 turnover, and in neutrophils simultaneously delay apoptosis (18, 19). However, more recently, caspase-dependent cleavage of Mcl-1 has been reported in a number of cell lines undergoing apoptosis triggered by agents such as staurosporine, Fas, and etoposide (22, 23, 24). Two evolutionarily conserved aspartic acid residues (at residues 127 and 157) have been identified in the PEST domain of the protein (24). However, there are conflicting reports on the biological function of the caspase-cleaved products. On the one hand, Michels et al. (24) report that the cleaved form (residues 128–350) has proapoptotic activity, whereas Clohessy et al. (23) suggest that this cleaved fragment cannot induce apoptosis.

In neutrophils, there is no evidence for caspase-mediated cleavage of Mcl-1 during constitutive or GM-CSF-regulated apoptosis (19). However, the accelerated rate of Mcl-1 turnover seen after sodium salicylate treatment was partly blocked by the pancaspase inhibitor, Z-VAD-FMK. Previous work has shown that this inhibitor can exert some unusual effects on neutrophils, at high concentrations (>100 μM) enhancing TNF-induced neutrophil apoptosis (38), but effectively inhibits caspases at the concentrations used in this study. Our data thus indicate that complex mechanisms for regulating Mcl-1 turnover via different subcellular routes can exist in cells: the proteasome or the caspases. The relative activities of these two proteolytic mechanisms is thus very much dependent on the signals received by the cells and the activities of the signaling networks that regulate their functions. It is possible that sodium salicylate results in a posttranslational modification of Mcl-1 that blocks its targeting the proteasome, or else that sodium salicylate interferes with proteasomal function, thereby preventing Mcl-1 (and presumably other proteins) being degraded via this route. It is also possible, and indeed perhaps more likely, that the primary event that triggers salicylate-induced Mcl-1 turnover is caspase activation, and that the caspase-cleaved form of Mcl-1 is proapoptotic, thereby greatly accelerating the apoptosis pathway. Rapid activation of several caspases following sodium salicylate treatment would appear to support this idea. We detected caspase-cleaved forms of Mcl-1 on Western blots (at ∼28 kDa), but these did not accumulate. This indicates that, as for full-length Mcl-1, caspase-cleaved Mcl-1 is also subject to high rates of turnover in neutrophils.

We have previously shown that signals generated via PI3K/Akt and Erk following GM-CSF signaling are important to induce posttranslational modifications to Mcl-1 that increase its stability and thereby delay neutrophil apoptosis (19). Increased Mcl-1 stability is thus a key event for GM-CSF signaling of neutrophil survival. This Erk-dependent phosphorylation to enhance Mcl-1 stability occurs on Thr¹⁶³ (21) within the PEST region. In this report, we show that sodium salicylate completely abolishes Akt activation and greatly decreases Erk activation in response to GM-CSF signaling. By this mechanism, sodium salicylate would prevent Erk-mediated phosphorylation of Thr¹⁶³ on Mcl-1 and thus prevent GM-CSF from signaling increased stability (19). This mechanism may help explain, at least in part, how sodium salicylate interferes with the normally protective effect of GM-CSF on Mcl-1 stability and survival.

In contrast to its effects on GM-CSF-regulated Akt and Erk activity, sodium salicylate induced a rapid and sustained activation of p38-MAPK. This effect was seen within minutes of addition of sodium salicylate in the presence or absence of GM-CSF, with activity remaining high throughout the 5-h incubation period. Sodium salicylate has been reported to exert a similar activation of p38-MAPK in a number of different cell types (20, 25, 26, 29), an event that offers at least a partial explanation for sodium salicylate-induced apoptosis. Indeed, it has been reported that the p38-MAPK inhibitor, SB203580, may partly block sodium salicylate-induced apoptosis (39, 40). However, in eosinophils and Jurkat cells, inhibition of p38-MAPK is reported to accelerate constitutive apoptosis (41, 42). The role of p38-MAPK in the control of neutrophil apoptosis is controversial, with some reports suggesting that it generates a survival signal (e.g., Refs.43, 44, 45) and some completely contrasting reports suggesting that it triggers a death signal (46, 47). In view of the dramatic and sustained activation of p38-MAPK and the rapid acceleration of apoptosis seen upon addition of sodium salicylate to neutrophils, it is very difficult to imagine that activated p38-MAPK can generate a survival signal. Based on the experiments presented in this report and those described above, we support the notion that p38-MAPK generates, either directly or indirectly, a death-promoting signal in neutrophils, at least in sodium salicylate-treated cells. Recently, it has been reported that salicylate induces cell death in leukemic cell lines and in parallel induces p38-MAPK activity (48).

If p38-MAPK generates a death signal in neutrophils, then inhibition of this signal should block apoptosis. In some cell types, this has been reported (39, 40). However, in our experiments, apoptosis measured by morphology after 8-h incubation control (untreated) or sodium salicylate-treated cells was completely unaffected by the addition of SB202190 (Fig. 8). However, these commonly used “p38-MAPK” inhibitors only selectively inhibit the α and β isoforms of the enzyme, but have no inhibitory effect on the γ and δ isoforms (49). Neutrophils only express the α and δ isoforms of p38-MAPK (50), and so these inhibitors will have no effect of the activity of neutrophil p38δ-MAPK. Therefore, it is possible that these α and δ isoforms have different effects on neutrophil apoptosis. It is speculated that p38δ-MAPK generates death signals that accelerate neutrophil apoptosis, particularly in the presence of sodium salicylate. In Jurkat cells, it has been proposed that p38β inhibits apoptosis, whereas p38α induces apoptosis (42). It is also possible that the activities of other kinases such as Akt and Erk can block death signals generated from p38-MAPK, and so complex and subtle control of cell fate is possible.

There is emerging evidence of cross talk between different MAPK to regulate cell function. For example, p38α-MAPK may activate the phosphatase PP2A, which then blocks MKK4-mediated activation of Jnk (43). Our data suggest that Akt may down-regulate p38-MAPK activity in the presence of GM-CSF, because the PI3K inhibitor, LY294002, resulted in enhanced activation of p38-MAPK in the presence of GM-CSF (Fig. 7 A). It is also possible that p38-MAPK may have regulatory effects on Akt and Erk activities. Therefore, complex and highly regulated control of the activities of these pathways is possible, which may positively or negatively regulate Mcl-1 stability and turnover and thereby neutrophil apoptosis.

We have previously shown that okadaic acid results in enhanced Mcl-1 turnover and accelerated neutrophil apoptosis (19), effects remarkably similar to those of sodium salicylate. The finding that okadaic acid also results in an extensive and sustained activation of p38-MAPK in control or GM-CSF-treated cells also resembles the effect of sodium salicylate. It is thus possible that, in this experimental system, sodium salicylate acts in a similar way as okadaic acid, i.e., acting as a phosphatase inhibitor. Alternatively, it may be that the effects of sodium salicylate are indirect, e.g., salicylate acts on a target that prevents phosphatase activation.

In summary, we have shown that complex mechanisms exist in neutrophils to provide for rapid changes in the turnover rate of Mcl-1 to either enhance or accelerate its turnover rate and increase or decrease, respectively, the rate of neutrophil apoptosis. In view of the importance of Mcl-1 overexpression in malignant or diseased cells (51, 52, 53), manipulation of these signaling pathways to accelerate Mcl-1 turnover may have therapeutic potential.

The authors have no financial conflict of interest.