Mechanisms Involved in Spontaneous and Viscum album Agglutinin-I-Induced Human Neutrophil Apoptosis: Viscum album Agglutinin-I Accelerates the Loss of Antiapoptotic Mcl-1 Expression and the Degradation of Cytoskeletal Paxillin and Vimentin Proteins Via Caspases

Neutrophils are involved in the inflammation process and because clearance of apoptotic neutrophils by cells such as macrophages can lead to the resolution of inflammation (1, 2, 3), it is important to develop therapeutic strategies based on the activation of neutrophil apoptosis to reverse or attenuate an inflammatory response. In addition to inflammation, deregulation of normal cell turnover via modulation of apoptosis may lead to cancer or autoimmune diseases.

Extracts of mistletoe (Viscum album) have been widely used in adjuvant chemotherapy of human cancer. V. album agglutinin-I (VAA-I)³ is a 63 kDa galactoside-specific plant lectin that belongs to the family of type II ribosome-inactivating proteins including abrin, modeccin, and ricin. The VAA-I molecule consists of two distinct subunits, the A chain (29 kDa) and the B chain (34 kDa). The A chain confers the property of the protein synthesis inhibitor to the VAA-I molecule by acting as a ribosome-inactivating agent. This is due to RNA-glycosidase activity that inhibits N-glycosylation of a single adenine within a universally conserved GAGA sequence on the 28S rRNA (4, 5). The B chain allows the VAA-I molecule to bind to terminal galactoside residues on membranes of various cells. VAA-I was recently found to act as a potent immunomodulator by activating different cell types (6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16). However, its interaction with human neutrophils has received little attention. This is intriguing because these cells are potent effectors of inflammation and are becoming increasingly recognized for their roles in anti-tumor reactions (17).

We have recently documented that VAA-I is a potent human neutrophil agonist (16). In particular, we have demonstrated that this lectin is a potent inducer of neutrophil apoptosis acting via inhibition of de novo protein synthesis and via activation of caspases that fragment gelsolin, a microfilament-associated protein (MFAP). Factors that modulate apoptosis and the execution phase of apoptosis may vary from cell to cell. In neutrophils, the common caspase substrates such as poly(ADP-ribose) polymerase, the catalytic subunit of DNA-dependent protein kinase, the small ribonucleoprotein U1-70 kDa and the nuclear/mitotic apparatus protein are not detected (18, 19). Because of this, it is important to identify caspase substrates in these cells. Because we have recently demonstrated that gelsolin is fragmented by caspases in VAA-I-induced human neutrophils (16), we hypothesize that several other cytoskeletal proteins may be targets to caspases.

The present study was conducted to better elucidate the mechanisms involved during VAA-I-induced human neutrophil apoptosis in comparison with spontaneous apoptosis. We found that VAA-I alters mitochondrial permeability and increases intracellular reactive oxygen species (ROS), and that its ability to induce neutrophil apoptosis requires its internalization, a process not dependent on a cell surface receptor-mediated pathway. We also found that Mcl-1 is an important target of VAA-I and that caspases are involved in the degradation of cytoskeletal proteins in both spontaneous and VAA-I-induced neutrophil apoptosis. In particular, the MFAP paxillin and the intermediate filament vimentin protein are very important in VAA-I-induced neutrophil apoptosis, but vinculin and α- and β-tubulin are not.

The plant lectin VAA-I derived from V. album was isolated and purified as previously published (16). The caspase-1, -3, -4, and -7 inhibitor, N-benzyloxycarbonyl-V-A-D-O-methylfluoromethyl ketone (z-VAD-FMK), was purchased from Calbiochem (La Jolla, CA). The following mAbs to human cytoskeletal proteins were purchased from Sigma-Aldrich (St. Louis, MO): anti-gelsolin (clone GS-2C4), anti-paxillin (clone PXC-10), anti-α-tubulin (clone B-5-1-2), anti-β-tubulin (clone 2-28-33), anti-vimentin (clone Vim 13.2), and anti-vinculin (clone Vin-11-5).

PLB-985 and PLB-985 cells deficient in gp91^phox (X-CGD; Ref. 20) were cultured in RPMI 1640 supplemented with 10% FCS (Sigma-Aldrich), at a cell concentration not exceeding 1.5 × 10⁶ cells/ml. Neutrophils were isolated from the venous blood of healthy volunteers by dextran sedimentation followed by centrifugation over Ficoll-Hypaque (Pharmacia Biotech, Baie d’Urfie, Canada), as previously described (16, 21). Blood donations were obtained from informed and consenting individuals according to our institutionally approved procedures. Cell viability (>98%) was monitored by trypan blue exclusion and the purity (>98%) was verified by cytology from cytocentrifuged preparations colored by Diff-Quick staining (Baxter, Miami, FL; Refs. 16 and 21).

To evaluate mitochondrial membrane potential, cells (1 × 10⁶/ml) were incubated with the cationic dye JC-1 (5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolocarbocyanine iodide; Molecular Probes, Eugene, OR) at a concentration of 1 μM for 15 min at 37°C (22) or with the cationic fluorochrome DiOC₆ (Ref. 3 ; 3,3′-dihexyloxacarbocyanine iodide; Molecular Probes) at a concentration of 40 nM for 30 min at 37°C (22, 23). The cells were then washed twice with PBS and analyzed by flow cytometry. Cells treated with 50 nM valinomycin (Juro, Lucerne, Switzerland) for 15 min were used as a control. The FACSCaliber was set up to measure forward and side scatter, green fluorescence (525 nm band pass), red fluorescence (FL2; 590 nm band pass) for the change from JC-1 aggregate (red fluorescence) to monomer (green fluorescence), or DiOC6 green fluorescence (Ref. 3 ; FL1; 530 nm). All flow cytometric analyses (10,000 events) were performed using CellQuest analysis software.

To determine potential intracellular increase levels of ROS induced by VAA-I, cells (1 × 10⁶ cells/ml) were incubated for 3 h with 1 or 10 μg/ml VAA-I. Intracellular levels of ROS were detected with the use of 2′,7′-dichlorofluorescein diacetate (H₂DCFDA; Molecular Probes) essentially as previously described (24). In brief, after stimulation, cells were washed with PBS and stained with the nonfluorescent cell-permeable H₂DCFDA (5 μg/ml) for 15 min at 37°C. The H₂-dichlorofluorescein oxidizes rapidly to highly fluorescent dichlorofluorescein by ROS. As a positive control, the fluorescence intensity of cells pretreated with H₂DCFDA was measured in the presence of 30 μM H₂O₂.

Assessment of neutrophil apoptosis was performed essentially as previously described (16, 21). Briefly, freshly isolated human neutrophils (10 × 10⁶ cells/ml in RPMI 1640 supplemented with 10% autologous serum) were incubated for 24 h in the presence or absence of 1000 ng/ml VAA-I. To investigate the possible involvement of H₂O₂ during VAA-I-induced apoptosis, catalase (Sigma-Aldrich) was added in cultures. In other experiments, cells were pretreated for 30 min with the vesicular transport inhibitor brefeldin A (Sigma-Aldrich) or for 1 h with the mitochondrial stabilizer bongkrekic acid (BA; Calbiochem; Novabiochem, La Jolla, CA) before VAA-I treatment. Cytocentrifuged preparations of neutrophils were performed with a Cyto-tek centrifuge (Miles Scientific, Elkart, IN), as previously described (16, 21) and were stained with a Diff-Quick staining kit (Baxter) according to the manufacturer’s instructions. Cells were examined by light microscopy at a final ×400 magnification and apoptotic neutrophils were defined as cells containing one or more characteristically dark-stained pyknotic nuclei. Results were expressed as percentage of cells in apoptosis. PLB-985 and X-CGD cell apoptosis were assessed as above, but cells were incubated at 1 × 10⁶ cells/ml.

Neutrophils (10 × 10⁶ cells/ml RPMI 1640-HEPES-penicillin-streptomycin) were stimulated 24 h with buffer, GM-CSF (65 ng/ml) or VAA-I (1000 ng/ml), and were harvested and washed twice with cold PBS. Whole cell lysates were prepared as previously published (16, 25). Proteins (1 × 10⁶ cells/well) were separated using 10% SDS-polyacrylamide mini gels and transferred to nitrocellulose. Membranes were blocked overnight with 3% nonfat dry milk (Carnation, Don Mills, Ontario, Canada) at 4°C in TBST (25 mM Tris-HCl (pH 7.8), 190 mM NaCl, 0.15% Tween 20). Mcl-1 Ab (K-20 clone; Santa Cruz Biotechnology, Santa Cruz, CA) was added at a final dilution of 1/200 in TBST + 3% nonfat dry milk for 1 h at room temperature. Membranes were then washed with TBST and incubated for 1 h at room temperature with a goat anti-rabbit-HRP secondary Ab (Jackson ImmunoResearch Laboratories, Mississauga, Canada) at 1/20,000 in TBST + 3% nonfat dry milk followed by washes. The Mcl-1 protein was revealed with ECL and quantified using a Fluor-S MultiImager (Bio-Rad, Hercules, CA) and the MultiAnalyst version 1.1 program (Bio-Rad).

Freshly isolated cells were incubated in the presence or absence of 1000 ng/ml VAA-I for 24 h as for the apoptosis assay. Cells were then washed and the cell concentration was adjusted to 1.5 × 10⁶ cells/150 μl and simultaneously permeabilized and fixed for 5 min with a mixture of 0.05% digitonin and 1.9% paraformaldehyde, a mixture known to preserve cytoskeletal proteins and to allow excellent staining (26, 27). Immediately after this incubation, all the corresponding tubes were filled with 4 ml of ice-cold PBS and washed twice. Cell pellets were suspended in 100 μl and incubated for 30 min at 4°C with a specific anti-human cytoskeletal protein mAb at a final concentration of 1/50. Cells were then washed twice and incubated with FITC-goat anti-mouse IgG Ab (1/50) for an additional 30 min at 4°C, light protected. Cells were then washed and analyzed by flow cytometry (10,000 events) using a FACScan flow cytometer (BD Biosciences, Mountain View, CA). In preliminary experiments, cells were preincubated for 30 min with 20% autologous serum to prevent any possible nonspecific binding via FcR. We obtained similar results whether cells were or were not pretreated with 20% autologous serum (data not shown). This reinforces the choice of our fixation/permeabilization procedure used to ensure suitable intracellular staining of cytoskeletal proteins. Because of this, we then omitted this step for all other experiments. The mean fluorescence intensity (MFI) obtained for a particular cytoskeletal protein from VAA-I treated cells was subtracted from the MFI obtained from untreated control cells (from which the MFI obtained with the conjugated alone was deduced). Results are expressed as a decrease of fluorescence (MFI-untreated − MFI-treated cells).

Neutrophils (10 × 10⁶ cells/ml in a 24-well plate) were incubated with or without 1000 ng/ml VAA-I in the presence or absence of z-VAD-FMK or its diluent (<1% DMSO) for the indicated period and then harvested for the preparation of cell lysates in Laemmli’s sample buffer. In some experiments, cells were preincubated for 1 h with the diluent or z-VAD-FMK before an additional 23 h of incubation (total period of 24 h). Results were similar whether cells were or were not pretreated for 1 h with z-VAD-FMK (data not shown). Therefore, in subsequent analyses, we omitted the preincubation step. Aliquots corresponding to 225,000 cells were loaded onto 10% SDS-PAGE and transferred from gel to polyvinylidene difluoride membranes (17). Nonspecific sites were blocked with 1% BSA in TBST overnight at 4°C. Membranes were incubated with monoclonal anti-human cytoskeletal Abs (anti-gelsolin, 1/1,500; anti-vimentin, 1/1,000; anti-vinculin, 1:150; anti-paxillin, 1/500; anti-α-tubulin, 1/500; or anti-β-tubulin, 1:500), for 1 h at room temperature, followed by washes, and incubated with a HRP-labeled sheep anti-mouse IgG (1/20,000; BIO/CAN, Montreal, Canada) for 1 h at room temperature in fresh blocking solution. Membranes were washed three times with TBST and bands were revealed with the ECL-Western blotting detection system (Pharmacia Biotech). Protein loading was verified by staining the membranes with Coomassie blue at the end of the experiments.

Statistical analysis was performed with SigmaStat for Windows Version 2.0 with a one-way ANOVA. Statistical significance was established at p < 0.05.

Fig. 1 illustrates that VAA-I alters the mitochondrial membrane because a depolarization was observed by flow cytometry using the JC-1 dye. We have also observed a VAA-I-induced depolarization using the dye DiOC₆ (Ref. 3 ; data not shown), but its use was recently found to be more reliable for analysis of plasma membrane potential rather than mitochondrial potential. As expected, valinomycin was found to markedly induce a depolarization of mitochondrial membrane. No major changes were observed with the use of a low VAA-I concentration (1–100 ng/ml; data not shown), a condition that does not affect neutrophil apoptosis (16). We have previously reported that VAA-I induces apoptosis in virtually all cells after a 24-h incubation period with a concentration of 1000 ng/ml (16). In this study, even at a concentration of 10 μg/ml VAA-I, we did not observe a greater mitochondrial membrane depolarization (Fig. 1). Neutrophils were incubated for 6 h in these experiments, based on our previous FITC-annexin V binding results (16).

BA is known to stabilize the mitochondrial membrane, preventing its disruption. This product was recently found to inhibit anti-Fas-induced neutrophil apoptosis when cells were preincubated for 1 h with 62 μM of BA (28). We have used BA under the same conditions and found that it cannot reverse the effect of VAA-I (Fig. 2). In two separate experiments, we observed that the anti-Fas Ab-induced apoptosis was reduced from 65 to 38% and from 75 to 62% with the use of BA (Fig. 2, inset). This correlates well with data previously reported (28).

Because ROS are known to be involved in neutrophil apoptosis (29, 30), we then decided to verify whether VAA-I can increase ROS. As illustrated in Fig. 3, VAA-I can increase intracellular ROS in human neutrophils. However, this was not observed with the lowest VAA-I concentrations (data not shown), correlating again with the lack of induction of apoptosis at lower concentrations (16).

Although we found that VAA-I increased intracellular ROS production, we had previously failed to detect extracellular superoxide (O₂⁻) production in VAA-I-induced neutrophils as assessed by a colorimetric method based on the reduction of ferrocytochrome c (V. Lavastre, unpublished observation). Therefore, we were interested in determining whether intracellular ROS are involved in VAA-I-induced neutrophil apoptosis which we studied by adding catalase to the cultures. Catalase is an inhibitor known to degrade H₂O₂ that is rapidly generated by superoxide dismutase when O₂⁻ is produced. As illustrated in Fig. 4 A, addition of a nontoxic catalase concentration (2000 U/ml) did not prevent VAA-I-induced apoptosis. In fact, the VAA-I-induced response was completely unaffected. This is in contrast to toxaphene (Sigma-Aldrich), a chlorinated hydrocarbon insecticide, that we recently identified as a potent inducer of human neutrophil apoptosis and a very potent activator of O₂⁻ production detected colorimetrically as described above (31). Note that, as expected, the addition of catalase in normal neutrophils undergoing spontaneous apoptosis was also diminished. This is in agreement with others (29). The addition of catalase, up to 5000 U/ml, did not cause any reversible effects (data not shown).

We then examined whether or not VAA-I could induce apoptosis in PLB-985 and X-CGD cells. As illustrated in Fig. 4 B, VAA-I induced apoptosis in both types of cells as assessed by cytology and by the degradation of paxillin, known to be fragmented in human neutrophils (in this study). This reinforces our conclusion that VAA-I does not induce apoptosis via ROS production.

VAA-I is a lectin known to be internalized which inhibits protein synthesis. Because of this, we next decided to use an inhibitor of vesicular transport, brefeldin A, and try to inhibit VAA-I-induced apoptosis. As illustrated in Fig. 5, pretreatment of cells with brefeldin A inhibited the ability of VAA-I, but not anti-Fas, to induce neutrophil apoptosis. This indicates that VAA-I needs to be internalized to mediate its effect and that induction of apoptosis by this lectin is not dependent on a cell surface receptor-mediated pathway, but rather is linked to its ability to inhibit protein synthesis, because the lectin needs to be internalized to dissociate, leading to ligation of the A chain to the ribosome.

To better understand how VAA-I induces neutrophil apoptosis, we then investigated the influence of this lectin on the expression of the recently identified new Bcl-2 family member, Mcl-1. This protein is known to protect against apoptosis and was recently found, unlike Bcl-2, to be expressed in human neutrophils (32, 33). As illustrated in Fig. 6, VAA-I down-regulates Mcl-1 expression in human neutrophils. As expected, GM-CSF, known to delay apoptosis, was found to prevent the loss of Mcl-1 protein when compared with control cells (33).

We recently found that the MFAP gelsolin is degraded following VAA-I treatment, via a caspase-dependent mechanism (16). Kothakota et al. (34) have clearly identified that the fragmentation of gelsolin is caspase-3-dependent. Because VAA-I is known to modulate different neutrophil responses involving cytoskeleton rearrangements, such as those which occur during apoptosis, we decided to study whether other cytoskeletal proteins can be targets of VAA-I. We first evaluated the protein content by a flow cytometric approach to answer whether or not VAA-I alters the expression of a particular protein in comparison with normal neutrophils undergoing spontaneous or constitutive apoptosis. Proteins belonging to the MFAP (gelsolin, paxillin, and vinculin), intermediate filament (vimentin), or microtubule (α- and β-tubulin) families were selected in the present study based on their importance in neutrophil cell physiology (35, 36, 37, 38, 39, 40, 41, 42). As illustrated in Fig. 7, we found that VAA-I-induced neutrophils express less gelsolin, paxillin, and vimentin when compared with 24 h-aged cells (or spontaneous apoptosis). Vinculin and both α- and β-tubulin protein expression did not differ between the two treatments. Our results demonstrate that VAA-I induces the degradation of gelsolin, paxillin, and vimentin but not vinculin or α- and β-tubulin.

We then confirmed the above flow cytometric results by studying the degradation of the cytoskeletal proteins as well as the possible involvement of caspases in this response. As illustrated in Fig. 8,A, paxillin was degraded by VAA-I over time. The corresponding Coomassie blue-stained membrane, shown in Fig. 8 (bottom panel), indicates that equivalent amounts of protein were loaded, reinforcing the disappearance of paxillin. The paxillin Ab used is known to react with at least two paxillin polypeptides with molecular masses of 68 and 40 kDa (43, 44). Our results are in agreement with this and VAA-I seems to preferentially degrade the bottom 40 kDa of polypeptide. This latter was only barely detectable after 18 h despite equivalent protein loading (Coomassie blue staining). However, note that we also detected two other polypeptides between the 68 and the 40 kDa bands. This may explain why others have detected a smear rather than the two distinct bands (43, 44). Interestingly, an unidentified ∼10 kDa fragment was detected over time, with maximal staining after 18 h, independent of the VAA-I treatment. This is probably a paxillin degradation product, because it was not detected by the other anti-cytoskeletal Abs used in this study. Curiously, this fragment was still detected when z-VAD-FMK was added to the cultures, suggesting that its fragmentation does not originate from the activity of caspases-1, -3, -4, or -7. However, the addition of z-VAD-FMK prevents the degradation of paxillin (68 and 40 kDa bands and those between) in a concentration-dependent manner (Fig. 8,B). The signal intensity was re-established near its initial value (Fig. 8,B, lane f, vs Fig. 8 A, lane 2). Our results indicate that paxillin, as gelsolin (16), is another MFAP degraded by VAA-I via caspases.

Using a similar approach, we found that, unlike paxillin, vinculin was not degraded by VAA-I treatment when compared with normal 24 h-aged neutrophils, despite the fact that ∼100% of cells were apoptotic after VAA-I treatment (Fig. 8,C). The addition of z-VAD-FMK to the culture prevents the loss of vinculin expression and re-establishes its expression to near the levels observed in fresh neutrophils (Fig. 8,D). These results are in agreement with the flow cytometric data obtained (Fig. 7).

We were then interested in answering whether or not the two major microtubule proteins, α- and β-tubulin were targets of VAA-I. As illustrated in Fig. 9,A, α-tubulin expression decreases over time at similar rate, whether cells were or were not treated with VAA-I. Similar results were obtained for β-tubulin (data not shown). Addition of z-VAD-FMK did not prevent α-tubulin degradation under any conditions (Fig. 9 B), and this was also observed with β-tubulin (data not shown). Our results indicate that, with respect to α- and β-tubulin, microtubules are not an important target of VAA-I and their progressively decreasing expression is not caspase-dependent. This is in agreement with the results of Atencia et al. (45), who reported that cleavage of tubulin is independent of ICE-like proteases. Again, our immunoblot results correlate well with those obtained using the flow cytometric approach.

We also investigated the role of intermediate filaments in spontaneous and VAA-I-induced neutrophil apoptosis. In human neutrophils, intermediate filaments are of vimentin type (39, 40). This explains why we decided to monitor potential degradation of this protein. As illustrated in Fig. 9,C, vimentin expression was markedly decreased over time by VAA-I treatment and its expression was undetectable after 24 h. Addition of z-VAD-FMK in the culture prevents vimentin degradation, attesting to the important role of caspases during this process (Fig. 9,D). These results correlate well with the marked differences in protein content detected by flow cytometry between spontaneous and VAA-I-induced neutrophil apoptosis (Fig. 7).

This study was performed to better understand the mechanisms by which VAA-I can induce neutrophil apoptosis. In addition, because it was previously reported that the MFAP gelsolin was found to be fragmented by caspase-3 during neutrophil apoptosis (34) and because we have recently found that this phenomenon also occurs in VAA-I-induced human neutrophils (16), we were interested in investigating the role of caspases in the degradation of other cytoskeletal proteins during both VAA-I-induced and spontaneous apoptosis.

We found that VAA-I alters mitochondrial potential, induces ROS production, and the loss of Mcl-1 protein and that caspases are involved in the degradation of gelsolin, paxillin and vimentin during apoptosis. However, VAA-I did not accelerate the degradation of α-tubulin, β-tubulin, and vinculin when compared with neutrophils undergoing spontaneous apoptosis. Our results have identified two novel caspase substrates, namely, paxillin and vimentin, involved in VAA-I-induced neutrophil apoptosis. It is presently difficult to directly link the ability of VAA-I to induce neutrophil apoptosis with the classical cytochrome c release observed in other cell types, because there is currently no clear evidence that cytochrome c is released in apoptotic neutrophils. We failed to detect cytochrome c release in VAA-I-induced neutrophils using a commercially available ELISA kit (V. Lavastre, unpublished data). Recently, using an interesting model, cytochrome c was found to be released in apoptotic neutrophils, but these cells have to be pretreated at 15°C for several hours (46), a situation that is unlikely to occur under normal physiological conditions. Involvement of cytochrome c in neutrophil apoptosis is still unclear and requires further investigation. It is well accepted that the protooncogene Bcl-2 is thought to control mitochondrial permeability transition, allowing the release of cytochrome c. The fact that mature human neutrophils do not express Bcl-2 could also, in addition to the few numbers of mitochondria they report, partly explain the poor ability of these cells to release cytochrome c. In recent years, other interesting Bcl-2-related members have been discovered and their expression at the protein and/or gene level has been detected in various cell types, resulting in the classification of a Bcl-2 protein family. Among these members, some are inhibitors (Bcl-2, Bfl-1, Bcl-x_L, Bcl-w, Mcl-1), while others are inducers of apoptosis (Bcl-x_S, Bad, Bak, Bax, Bik) (32, 33, 47, 48). In studies of human neutrophils, there are still some ambiguities concerning the expression of some of these Bcl-2 members.

In neutrophil cell physiology, the role of the Bcl-2 family of proteins in regulating the apoptotic rate is gaining increasing attention. Among them, Mcl-1 was recently found to be expressed in human neutrophils and its expression declined as neutrophils underwent apoptosis (32, 33). Agents that promote neutrophil survival such as GM-CSF, sodium butyrate, IL-1β, and LPS were found to prevent the loss of Mcl-1 expression (33). In this study, we found that VAA-I accelerates the loss of Mcl-1 protein when compared with neutrophils undergoing spontaneous apoptosis. Yet, it remains difficult to conclude whether the changes in Mcl-1 reflect rates of apoptosis or an inhibition of protein synthesis or even both. In this respect, our results are in agreement with the fact that VAA-I, like other ribosome inactivating proteins, acts as an inhibitor of protein synthesis (5, 16). Mcl-1 appears to have a very short half-life due to the presence of PEST sequences and Arg:Arg motifs that favor its proteolysis (49). Recently, the antisense oligonucleotide strategy for reducing Mcl-1 availability in human neutrophils was used and Mcl-1 was found to be necessary for delay of spontaneous or constitutive apoptosis (32). This attests to the importance of this antiapoptotic protein in human neutrophil cell physiology. Levels of Mcl-1 were recently found to increase with hypoxia-induced inhibition of apoptosis (32). We found that the VAA-I molecule induced such a response when compared with spontaneous apoptosis. This indicates that monitoring Mcl-1 expression in neutrophils is certainly a good marker of apoptosis as a decrease correlates well with induction of apoptosis.

Neutrophils isolated from CD95-deficient mice (lacking Fas) were found to undergo spontaneous apoptosis at the same rate as neutrophils from wild-type mice, in contrast to a putative role of Fas during constitutive apoptosis (50, 51). In this study, we found that pretreatment of cells with brefeldin A, an inhibitor of vesicular transport, could inhibit the ability of VAA-I to induce neutrophil apoptosis. This indicates that VAA-I needs to be internalized for mediating its effect and that induction of apoptosis, by this lectin, is not dependent on a cell surface receptor-mediated pathway. This correlates well with our previous observation that VAA-I does not induce tyrosine phosphorylation (16), a situation that occurs after receptor binding. Recently, we found that in addition to tyrosine phosphorylation, VAA-I does not induce threonine or serine phosphorylation events in human neutrophils (52). Taken together, these results strongly suggest that VAA-I does not induce phosphorylation events to mediate its proapoptotic effect. The ability of VAA-I to induce apoptosis independent of receptor expression has also been reported by others, but only in leukemic T and B cell lines (53).

ROS are known to be involved in neutrophil apoptosis (29, 54). H₂O₂ was found to increase both spontaneous and Fas-mediated apoptosis in neutrophils isolated from patients with chronic granulomatous disease, a hereditary defect in ROS production (30). The role of intracellular H₂O₂-induced production by PMA or ionomycin in normal human neutrophils was correlated with induction of apoptosis, whereas the inability of fMLP to induce apoptosis was associated with the extracellular production of H₂O₂ (55). Although we found that VAA-I increases intracellular ROS production in human neutrophils, reduction of ROS by the addition of catalase in the culture did not attenuate the ability of VAA-I to induce apoptosis. This suggests that this lectin does not induce neutrophil apoptosis via intracellular ROS production. In addition, VAA-I was found to induce apoptosis in promyelocyte PLB-985 cells as well as in X-CGD cells, confirming that ROS are not involved. We hypothesize that VAA-I accelerates apoptosis by shutting down the synthesis of proteins, including the antiapoptotic Mcl-1.

Caspases are known to be crucial mediators of apoptosis by catalyzing the cleavage of an increasing number of proteins. Among them, some are cytoskeletal proteins such as α-fodrin, gelsolin, Gas2, αII- and βII-spectrin, lamin A, keratins, and vimentin (34, 56, 57, 58, 59, 60, 61). In human neutrophils, caspases represent a central mechanism for both spontaneous and Fas Ab-induced apoptosis. To date, in human neutrophils, only the MFAP gelsolin was found to be a substrate for caspases (34). Actin is known to be cleaved during spontaneous neutrophil apoptosis, but this protein was not found to be a caspase substrate (62). Similarly, we found that the degradation of vinculin and α- and β-tubulin is not accelerated by VAA-I and that caspases do not seem to play an important role in their degradation during spontaneous apoptosis. In contrast, in addition to gelsolin, we found that paxillin and vimentin are two other important targets for VAA-I and their degradation is completely blocked by z-VAD-FMK. This indicates that caspases are highly involved in this process. In addition, caspases were also found to be involved in paxillin and vimentin degradation during constitutive apoptosis. Our results demonstrate the existence of a certain selective process for cytoskeletal protein degradation by caspases. Recently, two distinct pathways leading to nuclear apoptosis, one being caspase-dependent and the other being caspase-independent, were established in mouse embryonic fibroblasts (63). In this study, we provide further evidence that the pathways involved during spontaneous and VAA-I-induced human neutrophil apoptosis are caspase-dependent. We believe that the degradation of vimentin is more altered than paxillin because, unlike MFAPs, vimentin expression cannot be compensated by other candidates of the intermediate filament because it is the only candidate expressed in neutrophils. In the same vein, intermediate filaments are probably less dynamic structures than microfilaments known to be composed with several proteins.

In summary, we found that VAA-I induces neutrophil apoptosis via a particular pathway. The lectin does not appear to induce phosphorylation events (16, 52) and this correlates well with the absence of a receptor-induced apoptosis pathway. The lectin needs to be internalized and its ability to inhibit protein synthesis is linked to the induction of apoptosis. This correlates well with the loss of Mcl-1 protein expression. Although the lectin induces a depolarization of the mitochondrial membrane and increases intracellular ROS production, it does not induce apoptosis via the mitochondria-cytochrome c release pathway. Also, the lectin uses the caspase-dependent pathway for induction of apoptosis and gelsolin, paxillin, and vimentin are three important caspase substrates. Finally, we can conclude that the lectin elicits apoptosis in a receptor-independent fashion, but appears to use many of the same pathways that are typically associated with receptor-mediated apoptosis.

The plant lectin VAA-I acts as a very potent inducer of neutrophil apoptosis and this offers an interesting potential therapeutic strategy in the treatment of inflammatory disorders. Our results allow a better understanding of its mode of action in inducing neutrophil apoptosis. In addition, our results help elucidate the role of the cytoskeleton and caspases during spontaneous apoptosis in neutrophils. This is of significance for both general biology and medicine.

We thank Dr. Yves Pommier (National Cancer Institute, Bethesda, MD) for donation of the PLB-985 and X-CGD cell lines (20 ) and Mary Gregory for reading the manuscript.