Ca²⁺ influx shutdown in neutrophils induced by Fas (CD95) cross-linking

Abstract

In neutrophils, as in most other cell types, Ca²⁺ signalling is important for a number of cellular activities. Although inositol(1,4,5)trisphosphate-mediated release of Ca²⁺ from intracellular stores is a necessary prelude, it is the Ca²⁺ influx that is responsible for many of the neutrophil responses. We report here that although elevations of cytosolic Ca²⁺ do not accompany Fas-mediated apoptosis in neutrophils, the Ca²⁺ influx component of the response to N-formyl-methionyl-leucyl-phenylalanine (FMLP) becomes selectively inactived as the neutrophils progress towards accelerated apoptosis induced by Fas (CD95) cross-linking. After 4 hr incubation at 37°, untreated neutrophils display an exaggerated Ca²⁺ influx phase in response to FMLP. This was absent in neutrophils that had been Fas-activated at the same time. No Ca²⁺ influx component was demonstrable by the removal of extracellular Ca²⁺ or by Ca²⁺ channel blockade with Ni²⁺ and no Mn²⁺ influx was detectable. The defect could not be attributed to a decrease in receptor sensitivity, receptor coupling or receptor number because the release of stored Ca²⁺ remained constant during incubation and was unaffected by Fas activation. Ca²⁺ influx became uncoupled from store release before detectable gross morphological changes or phosphatidyl serine externalization and was also insensitive to caspase 3 and 8 inhibitors. These results suggest a mechanism other than caspase-mediated proteolytic damage to components important for Ca²⁺ influx.

Introduction

In neutrophils, as in most other cell types, Ca²⁺ signalling is important for a number of cellular activities, including the generation of oxidants and the release of proteases.^1,2 The agonist-evoked Ca²⁺ signal is caused by inositol (1,4,5)trisphosphate (IP₃)-mediated release of Ca²⁺ from intracellular stores and the secondary sustained phase of the Ca²⁺ signal is the result of Ca²⁺ influx across the plasma membrane activated by depletion of intracellular Ca²⁺ stores.^1–3 While the initial Ca²⁺ store release is a necessary prelude, it is the Ca²⁺ influx that is responsible for many of the neutrophil responses.^1,2 Neutrophils are highly susceptible in vitro to rapid apoptosis by Fas (CD95) cross-linking.⁴ The role of the Ca²⁺-signalling pathway in Fas-mediated neutrophil apoptosis is unknown. We report here that elevations of cytosolic Ca²⁺ do not accompany Fas-mediated apoptosis in neutrophils, but that the Ca²⁺ influx becomes increasingly uncoupled from store release. Furthermore, this Ca²⁺ influx shutdown was insensitive to caspase 3 and 8 inhibitors showing that Fas (CD95) cross-linking causes Ca²⁺ influx shutdown in neutrophils by a caspase-independent mechanism.

Materials and methods

Neutrophil isolation

Neutrophils were isolated from the heparinized blood of healthy volunteers as previously described.⁵ Following dextran separation, hypotonic lysis of red cells and centrifugation through Ficoll–Paque, neutrophils were resuspended in Krebs buffer (12+0 mm NaCl, 4·9 mm KCl, 1·2+ mm KH₂₊PO₄, 1·2+ mm MgSO₄, 1·3 mm CaCl₂₊, 2+5 mm HEPES and 0·1% bovine serum albumin, adjusted to pH 7·4 with NaOH).

Simultaneous cytosolic free Ca²⁺ and phase contrast imaging

Neutrophils were loaded with the Ca²⁺ indicator, fura2+-AM as previously described⁶ or with fura2+-dextran by micro-injection⁷ and were allowed to adhere to glass coverslips maintained at 37° on a temperature- and CO₂₊-controlled microscope stage system. Two excitation wavelengths 340 nm and 380 nm were sequentially transmitted to an inverted microscope (Nikon Eclipse) with an oil immersion 100× objective using a rapid monochromator (Delta RAM, PTI, Surbiton, UK). Phase-contrast images were taken simultaneously under far-red illumination (690 nm) using an appropriate dichroic mirror and a red-sensitive CCD camera. The fluorescent images were collected using a CCD camera (IC100 PTI, Surbiton, UK) and 340/380 nm ratio images were calculated using imagemaster software (PTI, UK). For the longest time-course, either illumination was continuous while continuous data was collected from regions of interest which corresponded to individual cells at one or two section intervals, or the illumination was discontinuous over periods of 4–6 hr and images were collected at defined intervals. The latter approach minimized the damage caused to cells by continuous illumination over extended times but necessarily had poor time resolution. Both strategies were used and combined so that high time resolution data could be obtained during key events such as during the cross-linking of Fas antibody or when the cell under observation was becoming morphologically apoptotic.

Micro-injection of fura2+-dextran

The large molecular weight conjugate of fura2+, fura2+-dextran (Molecular Probes, Eugene, OR; molecular weight 10 000), was employed to measure cytosolic free Ca²⁺ concentration with reduced diffusion of the fura2–Ca²⁺ complex within the cell. The fura2+-dextran was micro-injected into neutrophils by the simple, previously described,⁷ lipid-assisted micro-injection technique (SLAM) using premade SLAM pipettes (Cell Engineering Ltd, Swansea, UK). The probe was dissolved in intracellular medium (KCl, 150 mm, HEPES, 2+5 mm, pH 7·0) to give a final concentration of 500 μm and loaded into a micropipette (tip diameter 0·5 μm). On contact of the micropipette with the neutrophil, the transfer of fura2+-dextran into the cell was monitored by an increase in fluorescence at 360 nm to give intracellular concentrations of fura2+-dextran of between 10 and 50 μm. After successful micro-injection, the neutrophils remain fully functional, able to undergo phagocytosis in response to challenge⁸ appear morphologically normal, and maintain low cytosolic Ca²⁺ (100 nm).⁷

Fas stimulation

Neutrophils (5 × 10⁶ cells/ml) were treated with either anti-human Fas monoclonal antibody (1 μg/ml, DX2+; Oncogene, Boston, MA) or control antibodies (Sigma-Aldrich, Gillingham, Dorset, UK) for 5 min at 37°. The cells were allowed to adhere to glass coverslips, which were mounted in specially prepared chambers on a microscope stage, or in a two-chambered coverglass system (Nalge Nunc, Naperville, IL). The medium was replaced with RPMI-1640, containing fetal calf serum (10%) and cross-linking antibody (1 μg/ml, RaM Ab, Sigma-Aldrich) was added. The cells were maintained at 37° either in an incubator or on the microscope stage whilst maintaining the 5% CO₂₊ gas phase. All images were taken using an oil immersion 100× objective.

Results

Induction of apoptosis by Fas cross-linking in absence of Ca²⁺ signal

Cross-linking the irrelevant antibody has no significant effect on the spontaneous rate of apoptosis (Fig. 1a). However, cross-linking anti-Fas antibody caused a dramatic acceleration in the apoptotic rate, from apoptosis being barely detectable at 4 hr to being found in 80% of neutrophils by 8 hr (Figs 1a,b). Cross-linking anti-Fas antibody (Fas activation) had no detectable immediate effect on cytosolic Ca²⁺, measured over the first 30 min (Fig. 1c). There was also no significant difference in the fura2+ signals from populations of Fas-triggered and untreated neutrophils over longer periods (up to 2+ hr). However, fura2+ leakage over this time period was very significant under both experimental conditions and prevented further Ca²⁺ monitoring (Fig. 1c). To extend the period over which cytosolic Ca²⁺ could be followed, fura2+-conjugated to high molecular weight dextran (10 000) was micro-injected into individual neutrophils (Fig. 1d). This indicator was retained within neutrophils for at least 6 hr, during the time when morphological evidence of apoptosis could be demonstrated. However, there were no significant Ca²⁺ changes observable over the complete time-course (Fig. 1e). It was therefore concluded that elevations of cytosolic Ca²⁺ were not triggered by Fas activation and that an elevation of cytosolic Ca²⁺ was not required for progression to apoptosis by this accelerated route.

Figure 1.

Ca²⁺ signalling induced by Fas activation. (a) The graph shows the acceleration towards morphological apoptosis induced by cross-linking anti-Fas antibodies on human circulating neutrophils. Morphological apoptosis was identified by the darkly stained, condensed nucleus and cytoplasmic vacuolation. The rate of apoptosis was significantly different from the untreated apoptotic rate shown here. (b) The images show typical apoptotic neutrophils 2+0 h without (control) and with Fas activation. The effect was similar whether neutrophils were adherent to glass or kept in suspension on plastic. (c) The traces show the changes in cytosolic free Ca²⁺ in a suspension cell population after Fas activation. The upper trace shows a close-up view of the lack of a detectable Ca²⁺ change immediately after Fas activation and the lower trace shows a complete time course over many hours. The elevation in fura2+ ratio observed between 1 and 6 hr was attributed to leakage of fura2+ into the extracellular medium containing Ca²⁺ (1·3 mm, 37°). (d) The images show a typical experiment in which fura2+-dextran was micro-injected into single neutrophils, the image on the left showing the moment of micro-injection and the image on the right the resultant intracellular fura2+-dextran image (excited at 380 nm). (e) The graph shows a typical cytosolic free Ca²⁺ monitored by fura2+-dextran over 4 hr in individual neutrophils (5% CO_2+, 37°), which was observed on at least three separate occasions. (f) The images show phase contrast images of fura2+ loaded neutrophils after Fas activation (40 min and 4 hr).

Changes in Ca²⁺ signalling during progression towards apoptosis

To investigate the integrity of the Ca²⁺ signalling pathway during progression to apoptosis, N-formyl-methionyl-leucyl-phenylalanine (FMLP; 100 nm) was used as a standard challenge. This stimulus acts on the seven conventional transmembrane receptors^1,2 and signals Ca²⁺ through the generation of IP₃ and capacitative Ca²⁺ influx.^2,3 The Ca²⁺ signal generated was initially the classic single-peaked response (as a result of release of Ca²⁺ from stores) followed by an exponential decay (as a result of ATP-requiring pumping of Ca²⁺ out of the cell) to a slightly elevated plateau set by the equilibrium between Ca²⁺ influx and Ca²⁺ pumping (Fig. 2a). However, in untreated neutrophils, the second phase of the Ca²⁺ response became increasingly pronounced, developing a characteristic ‘kyphotic’ pattern in more than 90% of neutrophils by 4 hr (Fig. 2e). This ‘kyphotic’ Ca²⁺ signal was dependent on Ca²⁺ influx and the effect on cytosolic free Ca²⁺ probably arose from decreased Ca²⁺ extrusion as a result of the depletion of cellular ATP in the ageing neutrophils, whose ATP production relies on anaerobic respiration.⁹ Indeed, under the conditions in the experiments (i.e. with fura2+ loading), ATP levels fell to 2+5% of the initial value.¹⁰ However, in Fas-triggered neutrophils, by 4 hr, more than 90% of neutrophils retained their responsiveness to FMLP, but their Ca²⁺ responses did not show any kyphotic effect. The initial phase of the Ca²⁺ signal, which is related to the release of Ca²⁺ from intracellular stores, remained unaffected and was not significantly different from that observed at earlier times or in the controls at any times (Fig. 2). Furthermore, the rate of Ca²⁺ release in this phase remained faster than the time resolution of the imaging system (0·65–1 seconds) in both Fas-treated and control neutrophils, demonstrating that the signalling for IP₃ generation remained unaffected. In some experiments, the initial Ca²⁺ signal also began to diminish in magnitude compared to that in neutrophils in the control group after 6 hr. These results show that the Ca²⁺ signalling pathway in Fas-treated cells becomes increasingly modified as the cells progress to apoptosis and is consistent with an early defect in coupling to Ca²⁺ influx.

Figure 2.

Changes in Ca²⁺ signalling during progress towards apoptosis. The traces show the cytosolic free Ca²⁺ change within individual neutrophils in response to FMLP (100 nm, added at the point indicated ‘FMLP’) after incubation at 37° for the times shown without (control) or with Fas activation. The results shown are from a typical experiment, with the responses from three individual neutrophils superimposed. Traces (a, c, e) show the responses of untreated neutrophils and traces (b, d, f) show the responses from Fas-activated neutrophils at 40 min, 2 hr and 4 hr. The magnitude is shown as fura-2+340 nm/380 nm ratio (R) with the calibrated signals for the controls in this example corresponding to approximately 300 nm Ca²⁺ for the signal at 40 min and over 1 μm for the signal at 4 hr. The effect was observed in more than 90% of Fas-treated neutrophils on at least 10 other occasions.

Selective shutdown of Ca²⁺ influx pathway

To confirm that the changes induced by ageing of neutrophils and by Fas ligation were the result of an effect on the Ca²⁺ influx pathway, the effects of the absence of extracellular Ca²⁺ and of Ca²⁺ channel blockade were examined. In the absence of extracellular Ca²⁺ (less than 10 nm, EGTA 1 mm), the differences between 4 hr control and 4 hr Fas-treated signals were abolished and there was no significant difference between the Ca²⁺ signals (Fig. 3). As, under these conditions, the only source of Ca²⁺ for the cytosolic Ca²⁺ signal was from within the cell, i.e. within Ca²⁺⁺ stores, this confirmed that the coupling of steps between receptor occupancy to Ca²⁺ store release were not affected by Fas ligation. As removal of the Ca²⁺ influx component removed the differential effect, these results also showed that in Fas-treated cells Ca²⁺ influx was uncoupled from the Ca²⁺ store release.

Figure 3.

Dependence of ‘apoptotic’ Ca²⁺ signals on extracellular Ca²⁺. The traces show the cytosolic free Ca²⁺ change within individual neutrophils in response to FMLP (1 μm, added at the pointed indicated ‘FMLP’) after incubation (> 4 hr, 37°) in the presence and absence of extracellular Ca²⁺. Traces (a and b) show the responses from untreated neutrophils and traces (c and d) from Fas-activated neutrophils. Traces (b and d) show the Ca²⁺ responses in the absence of extracellular Ca²⁺ (labelled EGTA). In all traces, ionomycin (2+ μm) was added at the point marked ‘iono’ which confirmed the absence of extracellular Ca²⁺. The small and transient changes in cytosolic free Ca²⁺ were observed after addition of ionomycin in the absence of extracellular Ca²⁺ and attributed to the release of Ca²⁺ from sites within the neutrophils.

This conclusion was further confirmed by the addition of Ni²⁺ (1 mm), an ion which blocks Ca²⁺ channels on neutrophils but does not enter the cytosol (where it would quench fura2+ fluorescence). Stimulating neutrophils in the presence of Ni²⁺ removed the kyphotic phase of the Ca²⁺ signal induced by FMLP stimulation in 4-hr aged control cells (Fig. 4). In contrast, the presence of extracellular Ni²⁺ had no significant effect on the shape or magnitude of the response to FMLP in Fas-activated neutrophils (Fig. 4). Furthermore, by adding Ni²⁺ to neutrophils which were stimulated with FMLP, there was a clear reduction in the Ca²⁺ signal in control cells, showing that the Ca²⁺ channels were open at that time (Fig. 5a). In contrast, the addition of Ni²⁺ had no effect on the Ca²⁺ signal in 4-hr Fas-treated cells when added at the time of the plateau phase (Fig. 5b). However, in Fas-treated neutrophils which still had a kyphotic Ca²⁺ signal, the addition of Ni²⁺ reduced the Ca²⁺ signal to the levels observed in the non-kyphotic cell. Fura2+ quenching by an influx of Mn²⁺ ions (1 mm) was also absent in Fas-treated neutrophils (Figs 5e,f). As the influx of Mn²⁺ was also inhibited by Ni²⁺, these results confirm that Ca²⁺ influx channels, which were also permeable to Mn²⁺ and blocked by Ni²⁺, that normally open after stimulation which FMLP remained closed in the Fas-treated neutrophils.

Figure 4.

Effect of blockade of Ca²⁺ influx signals by Ni²⁺. The traces show the cytosolic free Ca²⁺ change in response within individual neutrophils to FMLP (1 μm, added at the pointed indicated ‘FMLP’) after incubation (> 4 hr, 37°) in the presence and absence of the Ca²⁺ channel blocking ion Ni²⁺ (1 mm). (a) and (b) show the responses from untreated neutrophils and traces, (c and d) show those from Fas-activated neutrophils. Traces (b and d) show the Ca²⁺ responses in the presence of Ni²⁺.

Figure 5.

Absence of Ni²⁺-dependent Ca²⁺ influx or Mn²⁺ influx in Fas-treated neutrophils. The traces show the cytosolic free Ca²⁺ change within individual neutrophils in response to FMLP (1 μm, added at the pointed indicated ‘FMLP’) after incubation (3·5 hr, 37° after Fas activation). Ni²⁺(1 mm) was added at the point indicated on the bar accompanying each trace. The results show a typical experiment. The traces show the effect of the addition of Ni²⁺ (1 mm) on the FMLP-induced Ca²⁺ response from (a) untreated neutrophils after incubation (control) and (b) Fas-activated neutrophils, and trace (d) shows the data from traces (b and c) superimposed to demonstrate that Ca²⁺ channel blockade by Ni²⁺ converted the ‘kyphotic’ response to the Ca²⁺ influx shutdown state. (e, f) In the presence of Mn²⁺ (0·1 mm), quenching of the fura2+ signal monitored at 360 nm (I_360nm) was used as a monitor of Mn²⁺ influx. In untreated neutrophils at 4–6 hr, an increase in Mn²⁺ influx was triggered with the increase in cytosolic free Ca²⁺-stimulated by FMLP (e) .However, in neutrophil treatment with Fas, FMLP failed to cause an increased influx of Mn²⁺ despite triggering a cytosolic free Ca²⁺± signal (f) .In both examples, which were typical of at least four others, ionomycin (2+ µm, added at the point marked ‘iono’) produced a large influx of divalent cations giving the maximum fura-2+340/380 ratio (Ca²⁺) and the maximum fura-2+360 nm quenching (I₃₆₀) as calibration points.

The effect of inhibition of caspases on Ca²⁺ influx shutdown

One possible mechanism for the inability to signal Ca²⁺ influx was that a protein, important for Ca²⁺ influx, such as the Ca²⁺ channel itself, was inactivated by the proteolytic activity of caspases that have a known role in apoptosis. The effect of caspase inhibition was therefore investigated. Neutrophils were preincubated for 1 hr with either caspase-3 inhibitor II (100 μm, Z-DEVD-FMK, Calbiochem, UK), or the caspase-8 inhibitor II (100 μm, Z-IETD-FMK, Calbiochem, UK). Although these treatments have an effect on the later morphological changes, Ca²⁺ influx shutdown in Fas-treated cells was unaffected by either caspase inhibitor (Fig. 6). These results show that Ca²⁺ influx shutdown was not caspase-dependent and it was concluded to be unlikely to result from protease-mediated intracellular ‘damage’. The Ca²⁺ influx shutdown was also demonstrable much earlier than gross morphological or nuclear changes on phase contrast microscopy or externalization of phosphatidyl serine (PS) as detectable by annexin V binding (see also Fig. 1).

Figure 6.

Effect of caspase inhibition on apoptotic Ca²⁺⁺ influx shutdown. The traces show the cytosolic free Ca²⁺⁺ change within individual neutrophils in response to FMLP (1 μm, added at the pointed indicated ‘FMLP’) after incubation (> 4 hr, 37° after Fas activation), with and without inhibition of caspases. The traces show the data from a typical experiment. Traces (a and b) shows the response from untreated (control) and Fas-activated neutrophils, respectively. Traces (c and d) show the effect of caspase 8 (Z-IETD-FMK, 1 hr, 100 μm) and caspase 3 (Z-DEVD-FMK, 1 hr, 100 μm) inhibition, respectively, on the response from Fas-activated neutrophils. Neither inhibitor generated the ‘kyphotic’ response, which would indicate effective Ca²⁺⁺ channel opening.

Discussion

The data presented here show that during accelerated apoptosis of human circulating neutrophils induced by Fas (CD95) cross-linking, the Ca²⁺ influx pathway triggered by FMLP becomes increasingly ineffective, and was absent in Fas-activated neutrophils by 4 hr. The Ca²⁺ influx shut-down could not be attributed to a decrease in receptor sensitivity, receptor coupling or receptor number because the release of stored Ca²⁺ remained constant during incubation and was unaffected by Fas activation. It would be expected that uncoupling Ca²⁺ influx from activating stimuli may have a profound effect on the responsiveness of neutrophils. Not only would they be unable to mount potentially pathogenic actions, such as extracellular oxidant and protease liberation, but phagocytosis, which requires Ca²⁺ influx¹¹ would also be impeded. These effects may underlie the previously reported decreases in neutrophil responsiveness on approach to apoptosis.¹² Physiologically, it may be important that apoptotic neutrophils have reduced sensitivity to stimuli and it could be argued that such a reduction in sensitivity is in line with the ‘purpose’ of neutrophil apoptosis in reducing the number of responsive cells at an inflammatory site which are potentially pathogenic. As we also show that Ca²⁺ signalling shutdown is an early event en route to apoptosis and is not dependent on caspase activity, it unlikely to be the result of apoptotic morphological changes or PS externalization. If PS externalization is the key signal for phagocytosis of apoptotic neutrophils¹³ events which occur after this are unlikely to have significant physiological consequences. However, Ca²⁺ signalling shutdown is one of the earliest events detectable and so would render neutrophils less responsive before full apoptosis and phagocytic clearance.

The effect we report here is similar to that reported for T cells. In T cells, Lepple-Wienhues et al.¹⁴ have shown that Fas-stimulation of T cells also blocked Ca²⁺ influx triggered by ligation of CD3. As the effect of Fas activation on Ca²⁺ influx was prevented in T cells lacking in acidic sphingomyelinase, and was restored by transfection with the enzyme, it was concluded that this enzyme was important for the Ca²⁺ influx pathway.^14,15 Fas may not be the only member of the superfamily of tumour necrosis factor (TNF) receptors to inhibit Ca²⁺ channels in the plasma membrane through activation of acid sphingomyelinase. TNF receptor ligands TNF-α and NGF have also been reported to attenuate Ca²⁺ influx in thyroid and mast cells, respectively.^16,17 However, in T cells, the production of IP₃ triggered by ligation of CD3 may also be inhibited by Fas ligation.¹⁸ This could not account for the uncoupling effect we report here for neutrophils, as the release of stored Ca²⁺ by FMLP, which is totally dependent on IP₃ binding to stores and can be inhibited by intracellular heparin or antibodies to IP₃ receptors,³ was unaffected at times when Ca²⁺ influx was abolished (see Figs 2,·3, 4,·5). In T cells, it has also been suggested that there may be an effect on membrane potential, a negative membrane potential providing the electrical component for Ca²⁺ influx. It has been shown that activation of Fas or the addition of ceramide inhibits N-type K⁺ channels through activation of a protein kinase, p56^1ck.¹⁹ Inhibition of K⁺ channels would cause membrane depolarization and would consequently reduce the electrochemical gradient and inhibit Ca²⁺ influx. Such an effect is unlikely to be relevant to neutrophils, as the oxidase is electrogenic and would already exert a large electrochemical restraint on Ca²⁺ influx.²⁰

Although Ca²⁺ influx is common to a number of non-excitable cells, the mechanism by which it occurs has yet to be fully resolved and may involve different channels in different cell types.^21,22 At least two mechanisms have been suggested, one involving conformational coupling of IP₃ receptors with a Ca²⁺ channel located in the plasma membrane²³ and another involving a diffusible second messenger released in response to Ca²⁺ store emptying and which causes Ca²⁺ channel opening.^24,25 Although the molecular identity is unknown, the latter factor may be a small (mol. wt < 1000) phosphorylated molecule²⁵ a similar activity having been isolated from Ca²⁺ store depleted neutrophils.²⁶ Although the mechanism for neutrophil Ca²⁺ influx shutdown is unclear, the possibility exists that it may involve residue mitochondria found in neutrophils. In basophils, it has recently been shown that ‘energized’ mitochondria are crucial for coupling Ca²⁺ store release to Ca²⁺ influx.²⁷ As mitochondria are known to take-up Ca²⁺ from the cytosol and are often placed close to the endoplasmic reticulum where Ca²⁺ release is occurring, a possible new role for mitochondria is implicated as the generator of the signal for Ca²⁺ influx.²⁷ In the neutrophil, released Ca²⁺ may thus be taken up by nearby mitochondria having significant membrane potentials, with a generation of the signal for Ca²⁺ influx. Although neutrophils have few mitochondria, which do not play a role in ATP production, the mitochondria often have membrane potentials.²⁸ Furthermore, recently it has been shown that an early event in the progress towards spontaneous neutrophil apoptosis, which precedes externalization of PS, is loss of mitochondrial membrane potential.²⁸ This may thus provide the link to Ca²⁺ influx shutdown demonstrated here.