Caspase-Mediated Proteolysis and Activation of Protein Kinase Cδ Plays a Central Role in Neutrophil Apoptosis

The PKC inhibitors bisindolylmaleimide (GF109203X) and Gö6976 were obtained from Calbiochem (San Diego, CA). Caspase inhibitors BD-fmk, zIETD-fmk, and zVDVAD-fmk were from Enzyme Systems Products (Dublin, CA) The phosphoinositide (PI) 3-kinase inhibitor LY294002 was from Biomol (Plymouth Meeting, PA). Recombinant human (rh) G-CSF was donated by Chugai Pharmaceuticals (London, UK), rhGM-CSF by Behringwerke/Hoechst (Marburg, Germany) and rhIL-3 by Sandoz (Frimley Park, UK). rhIL-6 and stem cell factor were from Peprotech (Rocky Hill, NJ). PKC antibodies were as follows: polyclonal anti-PKCα (C-terminal) from Sigma (St Louis, MO); monoclonal PKCβ (recognizes both βI and βII), γ, ε, θ, and ζ monoclonal antibodies (MoAbs) were from Transduction Labs (Santa Cruz, CA) and polyclonal antibodies (C-terminal) to PKCβΙΙ, PKCδ, and PKCι were from Santa Cruz (Santa Cruz, CA). A polyclonal ERK2 antibody was from Santa Cruz. Agonistic anti-Fas/CD95 (clone CH11) was from UBI (Lake Placid, NY).

Neutrophils from normal volunteers were isolated by dextran sedimentation, Ficoll hypaque density gradient (Amersham Pharmacia, Uppsala, Sweden) and hypotonic lysis of contaminating red blood cells as described previously.37 Cells were finally resuspended in RPMI 1640 supplemented with 1% heat-inactivated fetal calf serum (FCS) and incubated at a density of 10⁶ cells per mL in 24-well tissue-culture plates. At stated time points, cells were washed once with phosphate-buffered saline (PBS) and used in apoptosis quantitation assays. Fluorescein isothiocyanate (FITC)-conjugated annexin V (Boehringer Mannheim, Mannheim, Germany) was used according to the manufacturer’s instructions. Analysis of annexin V binding was performed by flow cytometry (Epics; Coulter, Hialeah, FL)38 and peak fluorescence of positive cells was at least 2 logs greater than the negative fraction.

Before cell lysis, neutrophils or HL-60 cells were washed once in PBS and incubated in 1 mmol/L diisopropylfluorophosphate (DIFP) for 5 minutes on ice. Cells were pelleted by centrifugation and resuspended in lysis buffer (50 mmol/L HEPES pH 7.5, 100 mmol/L NaCl, 1% Triton X-100, 1 mmol/L EDTA, 1 mmol/L EGTA, 20 mmol/L NaF, 1 mmol/L Na orthovanadate, 1 mmol/L Pefabloc, 10 μg/mL each of aprotinin, pepstatin, and leupeptin). After 5 minutes on ice, cells were centrifuged at 13,000g at 4°C for 10 minutes, assayed for protein concentration (Bio-Rad, Richmond, CA) and the supernatant added to boiling sample buffer and boiled for 10 minutes. Equivalent protein amounts were loaded for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on 10% acrylamide gels followed by transfer to nitrocellulose membranes (Hybond C; Amersham, Arlington Heights, IL). Primary antibodies were incubated with the nitrocellulose membrane for 60 minutes at room temperature at the following dilutions: anti-PKCα (C-terminal) 1:5,000; monoclonal PKCβ (recognizes both βI and βII) 1:1,000; polyclonal (C-terminal) PKCβΙΙ 1:1,000; anti-PKCζ 1:500; anti-PKCδ (C-terminal) 1:2,000; anti-PKCι 1:1,000; anti-ERK2 1:500.

Neutrophils were incubated ex vivo for 8 hours as described above. Cells were washed once in annexin V binding buffer (150 mmol/L NaCl, 5 mmol/L KCl, 1 mmol/L MgCl₂, 1.8 mmol/L CaCl₂, 10 mmol/L HEPES, pH 7.4) and resuspended in 90 μL of binding buffer plus 10 μL of annexin V microbeads (kind gift of Miltenyi Biotech, Bergisch Gladbach, Germany) per 10⁷ cells. Cells were incubated for 15 minutes at 12°C and then washed once in a large volume of binding buffer. Cells were loaded onto a mini-MACS column, washed with binding buffer, and both positively selected (by magnet) and flow-through cells collected. Cell lysates were made as described above, protein concentrations quantitated, and equal amounts of protein loaded for Western blotting. The purity of cells was verified by using annexin V-FITC and in the magnetically selected fraction 94% ± 3% of cells were annexin V positive and in the flow-through fraction 4% ± 2% were positive.

At appropriate time points, 5 × 10⁶ neutrophils were processed and Triton X-100 lysates prepared as described above. Immunoprecipitation of PKC isoforms was performed by the addition of 1 μg of a polyclonal anti–C-terminal antibody for 45 minutes and protein A-agarose for 45 minutes. Immunoprecipitates were washed twice with lysis buffer and once with kinase buffer (20 mmol/L Tris.Hcl, pH 7.5, 10 mmol/L MgCl₂). Samples were resuspended in a kinase mixture of 50 μL containing kinase buffer plus 20 μmol/L adenosine triphosphate (ATP), 5 μg H1 histone, and 2 μCi [γ-³²P]ATP (Amersham). Incubation was for 5 minutes at 30°C and was terminated by the addition of sample buffer and boiling for 10 minutes. Phosphorylated proteins were separated by SDS-PAGE, analyzed by autoradiography, and phosphorylation of substrate histone H1 quantitated using NIH Image software (NIH, Bethesda, MD). In experiments evaluating the effects of caspase inhibitors or agonistic Fas antibody on PKCδ activity, inhibitors were present throughout the cell culture period, but were not added further to the immunoprecipitated PKCδ. In experiments examining the effects of PKC inhibitors, GF109203X and Gö6976, these were present throughout the culture period at a concentration of 1 μmol/L; due to the competitive, reversible nature of these inhibitors, further compound was added to a final concentration of 500 nmol/L to the kinase mixture before the addition of ATP.

Purified recombinant caspases 3 and 8 and PKCδ were from Pharmingen (San Diego, CA). A total of 100 ng of PKCδ was incubated with 10 ng of caspase-3 or 150 ng of caspase-8 in a final volume of 50 μL of assay buffer (20 mmol/L 1,4 piperazinediethanesulfonic acid [PIPES], pH7.2, 100 mmol/L NaCl, 10 mmol/L dithiothreitol [DTT], 1 mmol/L EDTA, 0.1% 3-[(chlolamidopropyl) dimethylammonio]-1-propane sulfonate (CHAPS), 10% sucrose) for 2 hours at 37°C. Parallel tubes were incubated with 1 μmol/L AcDEVD-CHO (Bachem, Saffron Walden, UK) as a caspase inhibitor. The reaction was terminated by the addition of sample buffer and proteins were analyzed by SDS-PAGE and immunoblotting using the anti-PKCδ antibody.

Peripheral blood CD34⁺ cells were collected and purified from individuals with relapsed/resistant lymphomas undergoing stem cell mobilization for subsequent use after high-dose cytotoxic therapy. Cells were isolated and purified by immunoaffinity techniques as described. CD34⁺ cells surplus to defined clinical requirements were used for experimental purposes using a protocol with local ethics committee approval. A second round of selection using a mini-MACS column as per the manufacturer’s instructions was performed and yielded cells with 93% purity. Factor withdrawal-induced apoptosis was measured in freshly purified cells by incubation in a defined serum free medium (BIT; Stem Cell Technologies, Vancouver, Canada). To differentiate CD34⁺ cells, they were cultured in Iscove’s modified Dulbecco’s medium (IMDM)/20% FCS plus recombinant IL-3, IL-6, and stem cell factor (all at 20 ng/mL) for 7 days and then the same mixture supplemented with G-CSF (50 ng/mL) for a further 7 days. Factor withdrawal-induced apoptosis was measured after extensive washing and resuspension in serum-free medium. HL-60 and Jurkat cells were maintained in RPMI plus 10% fetal bovine serum.

Neutrophils were incubated in culture conditions as described above. At stated time points, cells were incubated in 5 μmol/L dichlorofluorescein diacetate (DCF-DA) for 60 minutes and then analyzed by flow cytometry. To control for variations in background fluorescence or dye uptake with time, parallel samples were incubated with the flavoprotein inhibitor diphenylene iodonium (DPI; Sigma) at 20 μmol/L for 15 minutes before DCF-DA addition. DPI inhibits oxidase activity, and in pilot experiments, inhibited the f-met-leu-phe–induced activation of the respiratory burst by over 94%. For each experimental condition, the DPI/DCF-DA sample was used as the control to which the DCF-DA alone sample was corrected. Hydrogen peroxide activity was expressed as the ratio of the mean channel fluorescence DCF-DA alone:DPI/DCF-DA. To verify that the addition of caspase inhibitors (BD.fmk, VDVAD.fmk, and IETD.fmk) did not directly affect the DCF-DA assay, they were incubated with neutrophils subsequently stimulated with phorbol ester to activate the respiratory burst. None of the caspase inhibitors reduced the response to phorbol ester in the DCF-DA assay.

Neutrophils undergo spontaneous apoptosis when incubated ex vivo (Fig 1) with survival of 57% ± 4% at 24 and 25% ± 5% at 48 hours. As has been reported previously, neutrophil survival is promoted by incubation with G-CSF.5Addition of the bisindolylmaleimide PKC inhibitor, GF109203X,39,40 also maintains viability with survival of 78% ± 3% and 65% ± 5% at 24 and 48 hours, respectively (Fig 1), suggesting that activation of PKC is causally involved in neutrophil apoptosis. To investigate which member(s) of the PKC family could be involved in this process, we examined the expression of PKC isoforms by immunoblotting with specific antibodies. Comparison of PKC expression in neutrophils was made with the myeloid cell line HL60, which can be induced to differentiate along the granulocytic pathway (Fig 2).41Of the classical PKCs, neutrophils expressed significant levels of PKCβ (using an antibody that recognizes both βI and βII isoforms), a low level of PKCα and PKCγ was not expressed in either HL60s or neutrophils. PKCδ was the only novel PKC isoform detected. Although HL60 cells express both ζ and ι/λ atypical PKCs, neutrophils express a low level of PKCζ only. These results are consistent with those published by other groups.42-44 

GF109203X has been shown to inhibit classical and novel PKCs, but to have no effect on the kinase activity of the atypical PKCζ.40,45 PKCζ activity is known to be regulated by 3-phosphorylated inositides generated by the action of phosphoinositide 3-kinase.35 Incubation with the PI 3-kinase inhibitor, LY294002,46 at 20 μmol/L did not attenuate neutrophil apoptosis (control 57% ± 6%, LY294002 60% ± 6% survival at 24 hours, n = 6), further suggesting that PKCζ is unlikely to play a significant role in neutrophil apoptosis. In control experiments, this concentration of LY294002 abolished the activation of the PI 3-kinase target Akt/protein kinase B in response to GM-CSF, and this inhibitory effect was maintained for at least 24 hours (data not shown).

Therefore, the antiapoptotic effects of GF109203X in neutrophils could be attributed to inhibition of the cPKCs α and/or β or of the nPKCδ. The PKC inhibitor, Gö6976, has been shown to inhibit classical PKCs, but to have a limited effect on PKCδ activity.45 47 Incubation with Gö6976 at concentrations up to 10 μmol/L did not inhibit neutrophil apoptosis (control 78% ± 2%, Gö6976 78% ± 3% survival at 24 hours, n = 3). Gö6976 (5 μmol/L) was functional under these conditions as shown by a 68% inhibition of phorbol ester-stimulated neutrophil hydrogen peroxide production as detected by the flow cytometric DCF assay. These results suggest that the effects of GF109203X on neutrophil survival are likely due to inhibition of PKCδ activity.

Induction of apoptosis in certain cell lines has been shown to result in the proteolytic cleavage of PKCδ leading to the generation of an active catalytic (C-terminal) fragment.18 Immunoblotting with an antibody raised to the C-terminal portion of PKCδ showed that a significant proportion of neutrophil PKCδ is proteolytically cleaved by 8 hours of incubation and that only the C-terminal fragment (and no full-length protein) can be detected at 20 hours (Fig 3). In contrast, using an anti–C-terminal antibody to PKCβII, which is the major neutrophil isoform,43 48 there was no detectable proteolysis of PKCβΙΙ between 0 and 20 hours of neutrophil incubation. Immunoblotting with an MoAb that recognizes both PKCβ isoforms showed no loss of full-length protein for up to 20 hours (data not shown). Immunoblotting with an anti–C-terminal PKCα antibody showed the presence of a lower molecular weight fragment, but this was present in freshly isolated neutrophils and did not alter with aging in culture. The degree of PKCδ proteolysis was not significantly affected by incubation with GF109203X.

Annexin V binding (a marker of apoptosis) indicated that only 6% ± 3% and 45% ± 5% of neutrophils were apoptotic at 8 and 20 hours, respectively, suggesting that PKCδ proteolysis precedes the onset of cell death. To verify this, annexin V–coated magnetic microspheres were used to separate neutrophils into annexin V negative and positive fractions and whole-cell lysates prepared for immunoblotting. Figure 4 shows that there is a significant proportion of cleaved PKCδ in the annexin V–negative cells and only the C-terminal fragment can be found in annexin V–positive cells.

In view of the detection of a C-terminal PKCδ fragment, which could have altered activity, kinase assays were performed on immunoprecipitated PKCδ. Figure 5 shows that there is spontaneous activation of PKCδ detectable at 8 hours of incubation. This activation can be blocked by incubation with the cPKC/nPKC inhibitor, GF109203X, but not with the cPKC inhibitor, Gö6976. In contrast, there was no increase in the activity of PKCα or PKCβ at 8 hours. The basal level of activity of PKCα and PKCβ was inhibited by Gö6976 confirming the activity of this inhibitor.

Previous studies have indicated a role for Fas/CD95 in neutrophil apoptosis. Incubation of neutrophils with an agonistic Fas antibody led to more rapid death compared with control cells (Fig 6) with over 30% of neutrophils undergoing apoptosis within 8 hours. Proteolysis of PKCδ was detectable by 2 to 4 hours of incubation with the anti-Fas antibody (Fig 6B), and there was a slight increase in PKCδ kinase activity at 8 hours compared with control cells (Fig 5).

The effect of cell permeable peptide inhibitors of caspase family members14 on PKCδ proteolysis and neutrophil survival was measured. Figure 7A shows that PKCδ proteolysis is markedly inhibited by the broad spectrum caspase inhibitor, boc.aspartyl.fmk (BD.fmk),14,49 and by the caspase-8 inhibitor, IETD.fmk,50 but not by another inhibitor designed to target caspase-2, zVDVAD.fmk.51Caspase-8 is recruited to the intracytoplasmic domain of oligomerized Fas/CD95 via its association with the adaptor protein FADD.11 Previous work has shown that PKCδ can be cleaved by the action of caspase-319; the caspase-8 inhibitor IETD.fmk could potentially block PKCδ proteolysis in intact cells by preventing the processing of caspase-3 by caspase-8. Alternatively, caspase-8 could also cleave PKCδ directly. We incubated purified active caspases 3 and 8 with purified recombinant PKCδ and found that only caspase-3 was capable of cleaving PKCδ, indicating that PKCδ is not a direct target of caspase-8 (Fig 7B). Neutrophil survival was enhanced in the presence of both BD.fmk and IETD.fmk (Figs 6 and 7C). VDVAD.fmk had no effect. IETD.fmk also inhibited anti-Fas–induced PKCδ proteolysis and apoptosis (Figs 6B and 7D).

Preincubation with GF109203X decreased the proapoptotic effects of anti-Fas antibody on neutrophils (Fig 7D). In contrast, incubation of Jurkat T-cells with GF109203X before the addition of anti-Fas antibody did not prevent apoptosis with survival of 24% in anti-Fas–treated cells and 11% in GF109203X+anti-Fas cells (measured at 24 hours). In Jurkat cells, PKCδ proteolysis was readily detectable by immunoblotting within 4 hours of Fas stimulation (data not shown). We further investigated the effect of GF109203X on apoptosis triggered by growth factor withdrawal in more primitive hematopoietic cells. GF109203X had no protective effect on factor-withdrawal–induced apoptosis in primary human CD34⁺ cells (survival at 24 hours in control cells 60%; +GF109203X, 58%). In addition, CD34⁺ cells were differentiated ex vivo in a growth factor mixture including G-CSF.52 After 14 days, when over 80% of cells were myelocytes and metamyelocytes, growth factors were removed. Under these conditions, PKC inhibition did not ameliorate apoptosis (control, 47%; GF109203X, 50%). These results indicate that the effect of GF109203X on cell death is restricted to terminally differentiated neutrophils.

There is increasing evidence that the production of ROS plays a critical role in neutrophil apoptosis; in particular,24neutrophils from patients with chronic granulomatous disease show prolonged neutrophil survival ex vivo. We used the DCF-DA fluorescent probe coupled with flow cytometry37 to look at the effects of GF109203X on neutrophil production of ROS. Figure 8 shows that there was a sustained increase in the spontaneous production of ROS in neutrophils incubated ex vivo and that this was inhibited by incubation with the PKC inhibitor, GF109203X. Incubation with the inhibitor, Gö6976, did not significantly affect ROS production (control, 2.1 ± 0.2-fold increase; Gö6976, 2.43 ± 0.1-fold; duplicate measurements in two separate experiments; mean ± standard deviation [SD]).

Incubation with the caspase-8 inhibitor, IETD.fmk, was also associated with a reduction in production of ROS, and this effect was not seen in cells incubated with the caspase-2 inhibitor, VDVAD.fmk, mirroring the effects of these inhibitors on neutrophil survival (Fig 7). In two separate experiments, incubation of neutrophils with the broad spectrum caspase inhibitor, BD.fmk, led to a decrease in ROS production compared with control cells (mean increase at 8 hours in control cells 2.1 ± 0.2 and in BD.fmk-treated cells 1.51 ± 0.1). Production of ROS was more pronounced in neutrophils incubated with an agonistic anti-Fas antibody (Fig 8). These results suggest that in this context ROS production is dependent on caspase activity, which is likely to be triggered via activation of the Fas/CD95 pathway.

In this study we have evaluated the role of PKC in neutrophil apoptosis and showed that the novel PKC isoform PKCδ plays an important part in mediating cell death. Neutrophils isolated from normal individuals and incubated ex vivo undergo constitutive apoptosis with about half of the cells dying in the first 24 hours. Other investigators have shown that this phenomenon can be inhibited by incubation with a variety of growth factors and cytokines.4-7 We show that incubation with the PKC inhibitor, GF109203X, also inhibits this form of cell death to levels comparable to cytokines such as G-CSF with only about 20% of cells undergoing apoptosis in the first 24 hours. Of the three major PKC isoenzyme groups, we show that neutrophils contain the classical PKCs α and β, the novel PKCδ, and the atypical PKCζ. This distribution of PKCs is similar to that reported by other groups.42-44 

The early development of PKC inhibitors was based on compounds such as staurosporine and H7, which act as catalytic domain inhibitors and are competitive with respect to ATP binding.40 Although these were effective, they lacked specificity. Staurosporine can, for example, inhibit many other types of kinases such as ribosomal S6 kinase, c-AMP–dependent kinase, and the src tyrosine kinase.40 The bisindolylmaleimide inhibitors are synthetic compounds based on staurosporine and have yielded potent and selective PKC inhibitors.39 However, these compounds vary in their potency against different PKC isoforms. The GF109203X compound used in this study has been shown to effectively inhibit the cPKCs α and β and the nPKCs δ and ε at concentrations in the nmol/L range; the 50% inhibitory concentration (IC50) for PKCζ is much higher at about 6 μmol/L.45 This specificity coupled with the distribution of PKC isoforms in neutrophils suggested that the effects of GF109203X on neutrophil survival could be a result of its effects on either the novel isoform PKCδ or the classical isoforms PKCα/β. We used the properties of another inhibitor, Gö6976,45 to discriminate between these possibilities. This inhibitor has been shown to inhibit the cPKC isoforms in the low nmol/L range, but does not affect the activity of PKCδ at concentrations up to 3 μmol/L.45 This failed to have any effect on neutrophil survival, suggesting that PKCδ was the most likely isoform responsible for the proapoptotic effects seen.

In 1995, Emoto et al showed that in U937 cells, induction of apoptosis in response to ionizing radiation was associated with proteolytic cleavage of PKCδ leading to the generation of an approximately 40 kD C-terminal fragment.18 This was catalytically activated due to the loss of the N-terminal inhibitory domain. Subsequent work from the same group showed that this cleavage (at aspartate 330) could be reproduced in vitro using purified caspase-3/CPP32.19 Transfection and transient overexpression of the PKCδ catalytic fragment could induce an apoptotic morphology in HeLa and NIH3T3 cells.19 These results suggested that PKCδ activation by proteolysis could be a participant in the apoptotic program and not just a consequence of the execution phase. Proteolytic activation of PKCδ has been implicated in several models of apoptosis. Caspase-3–dependent activation of PKCδ and PKCα in response to a variety of apoptotic stimuli was reported by Shao et al.53 Mizuno et al54 have shown that in human leukemic cell lines triggered to undergo Fas-mediated apoptosis, the nPKCs, including PKCδ, undergo limited proteolytic cleavage, and that this parallels the activation of caspase-3 and apoptosis. No data showing the effects of PKC inhibition were reported in this report, but in certain cell types, phorbol ester addition alone could induce apoptosis.

More recently, Denning et al55 have reported that in human keratinocytes UV radiation-induced apoptosis is associated with the proteolytic activation of PKCδ and that cell death can be inhibited by GF109203X preincubation. PKCδ proteolytic activation has also been implicated in cerebellar granule cell death.56 We have shown in this report that inhibition of PKC activity with GF109203X does not inhibit all forms of apoptosis, or indeed even all forms of apoptosis triggered by Fas ligation in different cell types. It has recently been shown that induction of apoptosis by Fas can lead to the cleavage of a number of signalling proteins.57 It is likely that caspase-mediated cleavage of different proteins may be critical for the apoptotic phenotype in various cell types and that this may be influenced by their extracellular environment. In neutrophils, cleavage and activation of PKCδ appears to play a crucial role in mediating apoptosis, but this may not be a rate-limiting step in other cell types.

Neutrophils undergoing spontaneous apoptosis show a rapid cleavage of PKCδ such that a significant proportion is cleaved at 8 hours and almost all by 20 hours. In contrast, there is no evidence of proteolysis of PKCα or PKCβ over this time course. The proteolytic changes in PKCδ are accompanied by increased enzyme activity and precede detectable features of apoptosis. At the 8-hour time point, only about 5% of cells display any features of apoptosis. One of the earliest markers of apoptosis is the externalization of plasma membrane phosphatidylserine (PS) detectable by the binding of annexin V.58-60 This PS exposure precedes the morphologic appearances of apoptosis, changes in membrane permeability to agents such as trypan blue and propidium iodide, or the characteristic DNA fragmentation. Functionally PS exposure is one of the mechanisms by which apoptotic cells are recognized by macrophages and targeted for ingestion.61 Separation of annexin V negative and positive neutrophil fractions using magnetic microspheres showed that annexin V negative cells contain a significant proportion of the catalytic PKCδ fragment. These results indicate that proteolytic activation of PKCδ takes place before the onset of apoptosis and that inhibition of PKCδ by GF109203X attenuates the apoptotic process.

Previous work has suggested a role for the Fas/CD95 pathway in neutrophil apoptosis.8,24 This pathway, which was initially thought to have a major role in T-cell regulation alone,9has been shown to be involved in apoptosis in several cell types in response to various inducers of cell death.62-64 Liles et al8 showed that spontaneous neutrophil death in vitro could be inhibited by blocking the Fas/CD95 ligand-receptor interaction and that neutrophils produced FasL suggesting an autocrine/paracrine mechanism of death. Neutrophil apoptosis can be accelerated by incubation with agonistic Fas/CD95 antibodies and this is associated with more rapid PKCδ proteolysis. After oligomerization of Fas/CD95 by ligand or MoAb, there is activation of caspase-8/FLICE, which is recruited to the receptor complex by FADD/MORT-1.11,12,50,65 Caspase-8 is capable of activating other effector caspases, such as caspase-3, by proteolytic removal of a prodomain.13 Thus caspase-8 sits at the apex of the Fas/CD95 pathway and its inhibition has been shown to prevent apoptosis in response to Fas/CD95 ligation.11,50 66 

We show that PKCδ proteolysis and activation is inhibited by preincubation with certain peptide caspase inhibitors. These are designed to mimic the preferred substrates of individual family members or to act as general inhibitors.14 The broad spectrum inhibitor, BD-fmk, decreases PKCδ proteolysis and apoptosis as might be expected. The inhibitor, IETD-fmk, targeted to caspase-8,50 also prevents both spontaneous and Fas/CD95-induced PKCδ proteolysis and apoptosis. Using purified recombinant enzymes, we show that caspase-3, but not caspase-8, can cleave PKCδ indicating that the effects of IETD-fmk are due to indirect inhibition of caspase-3 by preventing its activation by caspase-8 in the intact cell. These results confirm and extend those of Ghayur et al19 showing a major role of Fas/CD95 in spontaneous neutrophil apoptosis and indicate that PKCδ proteolysis plays a significant part in this process. The cleavage of PKCδ by caspase-3 in neutrophils where there are no other detectable features of apoptosis suggests that mechanisms to compartmentalize caspase-3 activation are likely to exist in the early stages of programmed cell death.

In addition to Fas/CD95, the neutrophil NADPH-oxidase system has been shown to play an important role in cell death. The production of ROS in other cell types can also predispose to apoptotic death.67Coxon et al25 showed that phagocytosis of opsonized particles by human neutrophils induced apoptosis in a CD11b/CD18 integrin and ROS-dependent manner. Kasahara et al24 showed that neutrophils from patients with CGD had much slower rates of constitutive or Fas/CD95-induced apoptosis compared with normals. Apoptosis of normal neutrophils can be inhibited by incubation with catalase, which leads to the rapid intracellular breakdown of hydrogen peroxide.22,24 We show that neutrophils incubated ex vivo produce increasing amounts of ROS with time and that this increase is inhibited by coincubation with either the PKC inhibitor, GF109203X, with the caspase-8 inhibitor, IETD-fmk, or the broad spectrum caspase inhibitor, BD.fmk. ROS production is not affected by incubation with Gö6976 or with the caspase-2 inhibitor, VDVAD.fmk. This suggests that activation of Fas/CD95 by endogenously produced Fas ligand promotes the formation of ROS in a caspase-8/PKC–dependent manner. Acceleration of the apoptotic process with activating anti-Fas antibodies leads to an increase in ROS production. Triggering the caspase cascade may also lead to the activation and/or dysregulation of other proteins involved in formation of ROS such as the Rac guanine nucleotide dissociation inhibitor D4-GDI68 and the p21-activated kinase PAK65.17 The latter may regulate the NADPH-oxidase complex by phosphorylating p47phox.69 The results shown here suggest that activation of PKCδ is crucial to the production of ROS in this system.

The precise role of PKC in the activation of the neutrophil oxidase system is not clear. The respiratory burst can be efficiently activated by PKC-activating phorbol esters or by synthetic analogues of diacylglycerol.70 These effects can be inhibited by competitive PKC inhibitors such as GF109203X.71 PKC may be involved in the phosphorylation and translocation of the p47 and p67 phox proteins, but the isoforms involved have not been defined.72-76 The data on the effects of PKC inhibitors on the respiratory burst induced by the bacterial peptide, fMLP, are contradictory, but this may relate to the use of nonselective inhibitors or to alternative pathways of oxidase activation.71,74,77-79 One group who used more selective inhibitors has suggested that chemotactic peptide-induced activation of the respiratory burst may depend on the activity of nPKCs,47 of which neutrophils only appear to express the δ isoform.

We have also shown that the effect of GF109203X on cell survival is dependent on the stage of differentiation. Primary peripheral blood-derived CD34⁺ cells and progeny induced to differentiate down the myeloid pathway in vitro undergo apoptosis on withdrawal of survival factors. This apoptotic pathway is not affected by the PKC inhibitor GF109203X, although CD34⁺ cells express several PKC isoforms including PKCδ (unpublished data); the only cells to be affected by PKC inhibition are terminally differentiated neutrophils. This may relate to the timing of development of a fully fledged NADPH-oxidase system, which occurs late in myeloid differentiation. In neutrophils, PKCδ acts as a link between the apoptosis-initiating Fas/CD95 receptor and the production of ROS, which amplifies the process of cell death. These results highlight the likelihood that although there are many conserved molecular themes in apoptosis, the precise interplay between component molecules may show significant variations even in cells of the same lineage.