A novel calcium-dependent proapoptotic effect of annexin 1 on human neutrophils

Abstract

The glucocorticoid-inducible protein annexin (ANXA) 1 is an anti-inflammatory mediator that down-regulates the host response. Endogenously, ANXA1 is released in large amounts from adherent polymorphonuclear neutrophils (PMN) and binds to their cell surface to inhibit their extravasation into inflamed tissues. The present study determined the effects of exogenous ANXA1 on several functions of human PMN in vitro. Addition of 0.1-1 microM human recombinant ANXA1 to the PMN provoked rapid and transient changes in intracellular Ca2+ concentrations that were blocked by the Ca2+ channel inhibitor SKF-96365. Although ANXA1 did not affect oxidant production and only minimally affected PMN chemotactic properties, the ANXA1-promoted Ca2+ influx was associated with two important functional effects: shedding of L-selectin and acceleration of PMN apoptosis. The latter effect was confirmed using three distinct technical procedures, namely, cell cycle, Hoechst staining, and ANXA5 binding assay. ANXA1-induced PMN apoptosis was insensitive to inhibitors of L-selectin shedding, whereas it appeared to be associated with dephosphorylation of the proapoptotic intracellular mediator BAD. In conclusion, exogenous ANXA1 displayed selective actions on human PMN. We propose that the new proapoptotic effect reported here may be part of the spectrum of ANXA1-mediated events involved in the resolution of acute inflammation.

Full text

The FASEB Journal express article10.1096/fj.02-0941fje. Published online June 17, 2003.
A novel calcium-dependent proapoptotic effect of annexin 1
on human neutrophils
Egle Solito,* Ahmad Kamal,† Francoise Russo-Marie,‡ Julia C. Buckingham,* Stefano Marullo,‡
and Mauro Perretti†
*Department of Neuroendocrinology, Imperial College London, Hammersmith Campus,
Commonwealth Building, Du Cane Road London W12 ONN, London UK; †The William Harvey
Research Institute, Bart’s and The London, Queen Mary SMD, Charterhouse Square, London
EC1M 6BQ, UK; ‡Department of Cell Biology, Institut Cochin de Génétique Moléculaire, 75014
Paris, France
Corresponding author: Egle Solito, Department of Neuroendocrinology, Imperial College
London, Hammersmith Campus, Commonwealth Building, Du Cane Road London W12 ONN,
London UK. E.mail: e.solito@ic.ac.uk
ABSTRACT
The glucocorticoid-inducible protein annexin (ANXA) 1 is an anti-inflammatory mediator that
down-regulates the host response. Endogenously, ANXA1 is released in large amounts from
adherent polymorphonuclear neutrophils (PMN) and binds to their cell surface to inhibit their
extravasation into inflamed tissues. The present study determined the effects of exogenous
ANXA1 on several functions of human PMN in vitro. Addition of 0.1–1 µM human recombinant
ANXA1 to the PMN provoked rapid and transient changes in intracellular Ca2+ concentrations
that were blocked by the Ca2+ channel inhibitor SKF-96365. Although ANXA1 did not affect
oxidant production and only minimally affected PMN chemotactic properties, the ANXA1-
promoted Ca2+ influx was associated with two important functional effects: shedding of L-
selectin and acceleration of PMN apoptosis. The latter effect was confirmed using three distinct
technical procedures, namely, cell cycle, Hoechst staining, and ANXA5 binding assay. ANXA1-
induced PMN apoptosis was insensitive to inhibitors of L-selectin shedding, whereas it appeared
to be associated with dephosphorylation of the proapoptotic intracellular mediator BAD. In
conclusion, exogenous ANXA1 displayed selective actions on human PMN. We propose that the
new proapoptotic effect reported here may be part of the spectrum of ANXA1-mediated events
involved in the resolution of acute inflammation.
Key words: inflammation • calcium channel • fMLP • chemotaxis • L-selectin
nnexins are a family of ubiquitous calcium and phospholipid binding proteins that are
present in all eukaryotes examined to date except Saccharomyces cerevisiae. These
proteins, of which 20 members are currently known, share a similar structure comprising
a core domain made of 70 amino acid repeats (usually 4 or 8 repeats) and a variable N-terminal
domain. The core domain confers both calcium and phospholipid binding properties (1), whereas
the N-terminal domain is believed to determine specific biological activities (1, 2).
A

Annexins are usually located intracellularly, either freely in the cytosol or associated with
membranes or cytoskeleton in a calcium-dependent manner. However, annexins are also secreted
and bound to the extracellular surface of the plasma membrane (3). Annexin 1 (ANXA1), a
protein particularly abundant in monocytes (4) and polymorphonuclear leukocytes (PMN) (5),
was originally identified as an endogenous mediator of the anti-inflammatory actions of
glucocorticoids (6). Glucocorticoids stimulate expression and/or secretion of ANXA1 in vitro (7)
and in vivo (8, 9). In experimental models of inflammation, administration of ANXA1 exerts
potent inhibitory effects on the process of cell migration (10, 11). Accordingly, passive
immunization against ANXA1 leads to a prolongation of the process of cell recruitment,
indicating that the endogenous protein promotes resolution of inflammation (12).
The mechanisms proposed for the anti-inflammatory effects of ANXA1 are multiple. ANXA1
was first shown to inhibit the activity of phospholipase A2, the enzyme that initiates the
biochemical cascade leading to eicosanoid’s synthesis (6, 13). A second mechanism by which
ANXA1 interferes with the inflammatory process was identified in leukocytes and was
correlated with their temporal and spatial activation. In resting PMN, most of the cytosolic
ANXA1 is stored in gelatinase granules (14, 15). Once PMNs adhere to endothelial cells, an
intense mobilization of ANXA1 on the cell surface occurs. The functional end point is a negative
regulation of the PMN trans-endothelial passage (16). Two recent studies have provided new
insight into the mechanism used by ANXA1 to block leukocyte extravasation. First, exogenous
ANXA1 was reported to interact with the α-4 integrin at the surface of myelocytic cells,
inhibiting its interaction with the endothelial integrin counter-receptor (17). Second, ANXA1
was shown to induce L-selectin shedding from the cell surface of human PMNs (18), preventing
their transmigration through the endothelium.
An interesting twist in this line of research has been the proposition that ANXA1-derived N-
terminal peptides, known to mimic some of the anti-inflammatory properties of the entire
protein, altered PMN transmigration in vitro through mechanisms involving the receptor for
formylated peptides (termed FPR) (19). In this study, putative FPR antagonists
[butyloxycarbonyl-, or Boc for short, derivatives of formyl-Met-Leu-Phe (fMLP) (20) abrogated
the effects of ANXA1-derived peptides on several assays of neutrophil activation (19). However,
Strausbaugh and Rosen did not find an involvement of FPR in full-length ANXA1-induced L-
selectin shedding from the cell surface of human PMNs (18). It is now clear that FPR is the
prototype of a complex family of receptors known to bind several chemically unrelated
molecules (21). The controversy that has evolved around the concept of ANXA1 acting at FPR is
somehow justified by the fact that so-called Boc antagonists do not discriminate among the
different members of the FPR family. More recently, we have provided evidence for an
interaction between endogenous and exogenous full-length ANXA1 and ANXA1-derived
peptides, with the lipoxin A4 receptor, a seven transmembrane G protein domain coupled
receptor structurally related to FPR (22).
Irrespective of the receptors involved, not many studies have investigated the effect of ANXA1
on neutrophil functions in vitro. A peptide drawn from the ANXA1 N terminus was shown to
inhibit neutrophil activation produced with interleukin-8 or platelet-activating factor (23), a
feature also demonstrated for the full-length protein (24). More recently, the ability of ANXA1-
derived peptides (19) and full-length ANXA1 (22) to cause transient changes in intracellular
calcium has been reported. The present study investigated the effects of human recombinant

ANXA1 on several markers of in vitro PMN activation, with particular emphasis on a novel
biological activity for this protein, that is, induction of a calcium-dependent early apoptotic
signal. As apoptotic PMN are rapidly removed by phagocytosis (25), a novel model emerges
where ANXA1 may limit PMN life span during inflammation.
MATERIALS AND METHODS
Materials
fMLP, bovine serum albumine (BSA), sodium tetraborate, ammonium formate, formic acid,
boric acid, lithium chloride, fura 2/acetoxymethylester (Fura 2/AM), 1-[β-[3-(4-
methoxyphenyl)propoxyl]-4-methoxyphenethyl]-1H-imidazole, HCl (SKF-96365, hydrocloride),
1,2-bis(o-aminophenoxy)ethan-N,N,N'N'-tetraacetic acid tetra(acetoxymethyl)ester (BAPTA),
and Hoechst H33342 were purchased from Sigma (St. Louis, MO). Boc1 (butyloxycarbonyl-
MDLP) was from ICN-flow (Emerald Biostructure, Bainbridge Island, WA). Dextran was from
Pharmacia Biotech AB (Uppsala, Sweden) and Lymphocytes Separation Medium was from
Eurobio (Les Ullis, France). Dowex AG1-X8 ion-exchange resin (200–400 mesh, formate form)
was from Bio-Rad (Hercules, CA), and myo(2-[3H])inositol (740 Gbq/mmol) was from ICN
Pharmaceuticals (Costa Mesa, CA). Hepes and Hank’s balanced salt solution (HBSS) were from
GIBCO (Life Science Cergy, Pontoise, France); the scintillation cocktail Hisafe3 was from
Pharmacia LKB (Bois d’Arcy, France). RO31-9790 and RO-28-2653 were a generous gift from
Dr. H.R. Krell (Roche Diagnostic GmbH Pharma Research, Penzberg Nonnenwald, Germany).
Human recombinant ANXA1, ANXA5, and chimeric ANXA1-5 (containing the first 44 aa of
ANXA1 attached to the 320-aa core of annexin 5) were produced in Escherichia coli and
purified by FPLC as already reported (11). Endotoxin contamination was <20 pg/ml as measured
by the Lymulus amebocyte chromogenic assay.
Human PMN isolation
PMNs were isolated from whole blood of healthy donors by a two-step procedure consisting of
dextran T500 (Pharmacia) sedimentation (45 min at room temperature) followed by
centrifugation at 1500 rpm for 20 min on Lymphoprep gradient (density 1.077 g/ml, Flobio,
France). Cells were used immediately after isolation. An enriched population containing >95%
PMN (red blood cell contamination was between 3 and 6% of the total cells) was routinely
prepared (26).
Functional studies on isolated PMNs
Changes in the cytosolic-free calcium concentration were measured in PMN loaded with 1 mM
Fura 2/AM at 37°C for 1 h (27). Fura 2 fluorescence assay was performed with aliquots of 5 ×
106 PMN in 2 ml HBSS, using a fluorimeter (Jobin Yvon 3D, Lonjumeau, France) equipped with
a thermally controlled cuvette holder and a magnetic stirrer. Excitation and emission wavelength
for Fura 2 fluorescence assays were 340 and 510 nm, respectively. Cytosolic calcium
concentrations were calculated as previously described (28). Tracings were reproduced and
scanned using an Agfa Snap CAM with version F-3.0 color software (Apple Macintosh).
Tracings are shown for one population of cells and are representative of at least three
experiments.

PMN (5×107/ml in HBSS supplemented with 0.025% BSA) (26) were labeled with 100 µM of
myo-(2-[3H])inositol/ml and incubated for 4 h at 37°C with gentle shaking. At the end of the
incubation period, cells were washed three times and resuspended in phosphate-buffered saline
(PBS) containing 20 mM Hepes, 0.025% BSA, and 10 mM LiCl (for inhibiting the enzymes of
InsP-dephosphorylation). After washing, cells were pooled again, counted, and resuspended in
20 mM H-HBSS, supplemented with 0.01% BSA and 10 mM LiCl. Cells (107/600 µl) were
aliquoted in Eppendorf and incubated for 10 min at 37°C in a water bath. The analysis of the
inositol was performed as described previously (29). After centrifugation, the hydrosoluble
perchloric supernatants were diluted and neutralized. The [3H]inositol-phospate was separated by
anion-exchange chromatography on Dowex AG1-X8 column (30). Radioactivity was determined
after adding 16 ml of scintillation cocktail to a 2-ml sample volume aliquot of each fraction.
Superoxide anion generation was measured by reduction of ferricytochrome C (horse heart type
III) as previously described (31). PMN (106 cells/ml) and 75 µl ferricytochrome C (5 mg/ml)
were incubated with or without recombinant ANXA1 for 10 min at 37°C. Incubation was
stopped by placing the tubes in an ice-water bath after which they were centrifuged at 1200 rpm
for 10 min. The absorbance of the supernatants was read at 550 nm in a spectrophotometer (DU
40 Beckmann, Gagny, France). The results were expressed in nmol of released O2–/min/106 cells.
A modified version of the original Boyden chamber technique was used for determining PMN
chemotaxis (32). In brief, 0.1 ml of PMN suspension (106 cells/ml) in HBSS supplemented with
1% bovine serum albumin (BSA) was added to the upper compartment of the Boyden chamber
and 0.2 ml of chemotactic substances to the lower compartment. A cellulose filter (Millipore
Rond-Point des Sangliers, Guyancourt, France) of pore diameter 3 µm was placed between the
two compartments. Following 90 min at 37°C chemotaxis in response to fMLP with or without
ANXA1 addition, filters were removed and stained with hematoxylin (Sigma, St Louis, MO).
Cell migration was determined by measuring the distance of the leading front within the filter
(33), using a light microscope and an oil immersion 100× objective. Five high-power fields were
analyzed for each filter by an observer unaware of the different treatments.
A whole blood protocol was used to minimize the risk of PMN activation. Venous blood was
collected in tubes containing anticoagulant from human volunteers. Blood aliquots (200 µl) were
treated with different stimuli at 37°C for 15 min. At the end of incubation, blood aliquots were
incubated for 10 min on ice, before erythrocyte lysis, which was produced by addition of an
equal volume of Immuno-Lyse solution (Coulter, Miami FL). A monoclonal antibody against
human L-selectin (CD62L; 20 µg/ml of clone MCA1076G, Serotec, Oxford, UK) was added for
1 h at 4°C, followed by a 30-min incubation with a FITC-labeled goat anti-mouse IgG (Serotec
1:200). After three washes, cells were analyzed by flow cytometry. Each leukocyte population
was identified by its forward and side scatter profile and by using specific cell markers (CD14
for monocytes and CD16 for neutrophils). Fluorescence in the FL1 (green) channel was
quantified as a measure of L-selectin expression vs. a background control obtained with cells
stained with FITC-mouse IgG alone.
Studies of PMN apoptosis
Apoptosis was determined by staining DNA with propidium iodide (PI), using a procedure
recently described in detail (34). In brief, resting or treated cells were washed twice with cold

PBS and fixed in cold 80% ethanol for 30 min. After they were washed in PBS, cells were
incubated in a PI buffer [PBS pH 7.4, 0.1 mM ethylenediaminetetraacetatic acid (EDTA), 0.1%
Triton X-100] containing 5% RNase T1, 1 mg/ml RNase A, and 100 µg/ml PI (final
concentrations). Distribution of cell cycle phases with different DNA content was determined by
flow cytometry. A pass filter of 585 nm was used to collect PI fluorescence, acquiring 10,000
events for each sample.
In some experiments, apoptosis was also measured by Hoechst nuclear staining (35). After
appropriate treatments, cells were resuspended in 50 µl at a concentration of 1 × 106 cells/ml,
and 5 µl of a solution of Hoechst H33342 (10 µg/ml) were added. Cells were then further
incubated for 20 min at 37°C and subsequently fixed with 10 µl of cold formaldehyde (40%
w:v). Cells were then analyzed by fluorescence microscopy with an Olympus BH-2-RFCA
microscope. Apoptosis was assessed on the basis of chromatin structure and condensation. For
each condition, at least 200 cells were counted.
Finally, to differentiate between early and late apoptosis, we quantified FITC-annexin 5 (FITC-
ANXA5) binding as recently described (34). In brief, 106/ml PMN were washed in cold PBS and
resuspended in 1× binding buffer in the presence of FITC-ANXA5 (5 µl; BD-Bioscience,
Oxford, UK) + 10 µl of PI (50 µg/ml) and incubated at room temperature for 15 min in the dark.
Analysis by flow cytometry was performed within 1 h using a FACScan II analyzer (Becton
Dickinson, Cowley, Oxford, UK), with 15 mW argon ion laser (488 nm) and Consort 32
computer running Lysis II software. FITC-ANXA5 staining was detected in the FL1 (green)
channel, whereas PI staining was monitored in the FL2 (red) channel. The following controls
were used to set up compensation and quadrants: unstained cells and cells stained with FITC-
ANXA5 or with PI alone. Other samples were also incubated with FITC-ANXA5 in buffer
without calcium to control for nonspecific binding.
Analysis of BAD phosphorylation
PMN (107) were incubated at 37°C for 30 min with human recombinant ANXA1 (0.5 µM) in the
presence or absence of SKF 96365 (25 µΜ). Cells were then lysed in 20 mM Tris (pH 7.5), 150
mM NaCl, 1% Triton X-100 containing aprotinin (0.15 U/ml), 20 mM leupeptin, and 1 mM
phenylmethylsulfonylfluoride (PMSF), 1 mM NaF, and 1 mM NaVO4. Immunoprecipitation
experiments were performed on cell lysates. In brief, 200 µl were incubated with rabbit
polyclonal anti-BAD [directed against a synthethetic peptide (KLH conjugated) corresponding to
Bad residues 112–130 (Cell Signaling, New England Biololabs, Beverly, MA); this antibody
does not discriminate between phosphorylated and nonphosphorylated BAD protein], with gentle
rocking overnight at 4°C. Protein A agarose beads were than added, and samples were further
incubated for 3 h at 4°C. Samples were washed in lysis buffer 5 times, and pellets were
resuspended in 3× SDS sample buffer and then applied on 12% SDS-PAGE and transferred to a
nitrocellulose membrane. The membranes were first blocked with 3% milk powder for 1 h,
followed by the incubation with the above anti-BAD polyclonal antibody (1:1000) or anti-serine
(Clone 4A9; BIOMOL, Plymouth Meeting, PA) monoclonal antibody (0.5 µg/ml) overnight. The
secondary antibody was an anti-rabbit or anti-mouse IgG-horseradish peroxidase conjugate (IgG-
HRPO, Sigma) and was used at a final dilution of 1:2500.

Statistical analysis
Experiments were performed in duplicate and triplicate and repeated at least three times to obtain
a mean ±SE. Differences between two experimental groups were analyzed by the Student’s t test,
whereas comparisons among three or more groups were made by ANOVA followed, if
significant, by Bonferroni post hoc test. P< 0.05 was considered significant.
RESULTS
ANXA1 produces Ca2+ fluxes in human PMN
Previous studies showed that exogenously added ANXA1 binds to the extracellular surface of
human neutrophils (4). We used a preparation of a recombinant purified protein (ANXA1),
produced in E. coli, shown to possess antimigratory activity in vivo (11), to study signaling and
functional effects associated with ANXA1 binding to human PMN.
ANXA1 added to human PMN produced transient changes in intracellular Ca2+ concentrations
([Ca2+]i) (Fig. 1A, 1B). The traces appeared slightly slower in their kinetics and dome-shaped
compared with those elicited by fMLP (10 nM; Fig. 1A). The effect of ANXA1 was
concentration-dependent, detectable at 200 nM and maximal at 1 µM (Fig. 1B). The homologous
protein ANXA5 produced a much weaker Ca2+ signal than ANXA1, whereas a chimeric that
contained the N-terminal domain of ANXA1 and the core domain of ANXA5 promoted a strong
Ca2+ signal, which was even larger in amplitude than that elicited by ANXA1 (Fig. 1C). These
data do not support the hypothesis that the putative Ca2+ transport activity of the annexin core
domain (36, 37) is involved in the Ca2+ signal observed with human PMN. Conversely, they
suggest a specific role for the unique ANXA1 N terminus. Coincubation with the calcium-entry
blocker SKF-96365 produced a concentration-dependent inhibition of ANXA1-induced [Ca2+]i
fluxes (Fig. 1D, 1E). Although this compound produced total inhibition of the response elicited
by ANXA1, it affected only the second phase of the response elicited by fMLP (Fig. 1D). In fact,
PMN activation by fMLP has long been characterized by an initial increase in [Ca2+]i levels by
opening intracellular stores, followed by a slow inward phase sustained by entry of the cation
from the extracellular medium through voltage-dependent calcium channels present in the
plasma membrane (38, 39).
Due to the recently proposed link between ANXA1 and FPR we compared, at least in part, the
calcium signals by fMLP with the one produced by ANXA1. Figure 2 summarizes this set of
data. Both fMLP- and ANXA1-induced [Ca2+]i responses were inhibited by the FPR antagonist
Boc1 (Fig. 2A, 2B). Cross-desensitization analysis showed that fMLP desensitized ANXA1-
induced [Ca2+]i response, whereas the opposite did not occur (Fig. 2C). This is identical to what
we have recently reported for the ANXA1-derived N terminus peptide (22).
Therefore, the Ca2+ responses produced by fMLP and ANXA1 in human PMN are quite different
despite being equally susceptible to the effect of the FPR antagonist Boc1. Based on these initial
analyses, we selected a concentration of 500 nM ANXA1 for most of the subsequent
experiments.

Other PMN responses
The diversity of effects between fMLP and ANXA1 was also evident with respect to classical
markers of PMN activation. In contrast to fMLP, ANXA1 failed to activate InsP formation (Fig.
3A) and superoxide release (Fig. 3B). In these assays, addition of ANXA1 alone to cells did not
produce any significant effect. Slightly different was the result obtained in the chemotaxis assay.
ANXA1 did not inhibit the marked chemotactic effect of fMLP (Fig. 3C); however, at 250 nM, it
produced by itself a modest yet significant chemotactic response.
Rapid cleavage of L-selectin from the cell surface is a well-established marker of PMN
activation. Using a whole blood protocol, we found that human recombinant ANXA1, but not
ANXA5, produced L-selectin shedding (Fig. 4A illustrates a representative experiment, with the
histograms displaying the fluorescence measured in the FL1 channel). In analogy to that
observed in terms of [Ca2+]I, the chimeric protein ANXA1/5 retained full activity. Similar data
were obtained for the monocytes (data not shown). Cumulative data of this series of experiments
is shown in Figure 4B.
Next we performed some mechanistic experiments using L-selectin expression as a read-out of
PMN activation. First, we linked ANXA1-induced L-selectin shedding to the changes seen in
[Ca2+]i (see Fig. 1 and 2). Addition of the calcium-entry blocker SKF-96365 abrogated ANXA1-
induced L-selectin shedding (Fig. 5A), whereas this compound was essentially inactive on the
fMLP response. Second, we used the FPR antagonist Boc1 (20 µM) that reverted the shedding
produced by either fMLP or ANXA1 (Fig. 5B). With either stimulus, L-selectin shedding was
genuinely provoked by the specific metalloprotease recently identified and characterized (40,
41). In fact, the selective inhibitor RO31-9790 used at a concentration of 50 µM, which retains
selectivity (18), but not the unrelated compound RO28-2653, a gelatinase inhibitor, prevented
fMLP- and ANXA1-induced L-selectin shedding (Fig. 5C).
ANXA1 induced human PMN apoptosis
Because changes in [Ca2+]i sometimes have been linked to induction of cell apoptosis, we next
investigated whether the anti-inflammatory mediator ANXA1 would accelerate PMN apoptosis.
Human PMN incubation with ANXA1 (6 h) resulted in a concentration-dependent acceleration
of PMN apoptosis as determined by FACS analysis, using the cell cycle/PI procedure. This
incubation time was chosen in preliminary experiments and was found to be optimal, because
longer incubations (e.g., overnight) resulted in a marked degree of spontaneous apoptosis (>50–
60%) and were therefore more difficult to modulate. Figure 6A shows a representative FACS
profile for PI staining in control and ANXA1-stimulated PMN. This novel property of ANXA1
was confirmed using another classical procedure to detect programmed cell death. Figure 6B
shows representative pictures of nuclear fragmentation as visualized with the Hoechst staining.
Finally, double staining with FITC-ANXA5 and PI also detected ANXA1-induced early (FITC-
ANXA5+, PI– cells) and late (FITC-ANXA5+, PI+) apoptosis (Fig. 6C). Out of three experiments
performed with different donors, PMN incubation with 500 nM ANXA1 produced a 6 h degree
of apoptosis that ranged between 15% and 50% (always significantly higher than the one
measured in untreated PMN, ranging from 3–5% to 25%, respectively).

Previous studies showed an increase in caspase-3 activity during constitutive apoptosis of aging
PMN and apoptosis induced by anti-Fas antibodies (42). In contrast, apoptosis induced by the
neutrophil activator PMA as well as tumor necrosis factor (TNF)-α is caspase-3 independent,
indicating that this cell type possesses at least two different modes for entering into programmed
cell death (42, 43). The same may apply to ANXA1, because caspase-3 activity after ANXA1
incubation (at 500 nM for 6 h) was not different from that measured in control cells (data not
shown). Moreover, PMA- and ANXA1-mediated apoptosis were similar in their extent and only
partially additive (Fig. 7A), suggesting the possible activation of a common pathway. In these
experiments, the effect of fMLP was also tested, finding induction of apoptosis (Fig. 7A).
However, more accurate analysis by the FITC-ANXA5 and PI double staining indicated that the
fMLP effect was delayed compared with ANXA1 (data not shown). PMN apoptosis was not
mimicked by ANXA5, whereas the chimeric protein ANXA1/5 produced again an effect that
was qualitatively and quantitatively similar to that of ANXA1 (Fig. 7B).
Consistently, ANXA1-induced apoptosis was abrogated by the blocking channel SKF 96365 in a
concentration-dependent manner (Fig. 8A). This blocker was ineffective on fMLP-induced PMN
apoptosis (data not shown). Furthermore, the calcium-chelating agent BAPTA attenuated the
effect of the protein, whereas the L-selectin sheddase inhibitor RO31-9790 was not active in this
experimental context (Fig. 8B).
Finally, PMN incubation with ANXA1 produced dephosphorylation of Bad, a classical
proapoptotic signal molecule (44). Figure 9A shows Bad immunoreactivity in PMN protein
extracts immunoprecipitated with a polyclonal antibody against Bad and blotted against a serine
antibody (Fig. 9A, top panel) or a polyclonal antibody against total Bad (Fig. 9A, bottom panel).
Out of three experiments, an approximate 40% reduction in phospho-BAD was calculated
following PMN incubation with ANXA1 (Fig. 9B). Again, these effects were susceptible to
inhibition by the SKF-96365 compound (Fig. 9A, 9B).
DISCUSSION
The main result of the present study is the novel notion that ANXA1-induced activation of
human PMN is associated with an accelerated degree of cell apoptosis. Addition of ANXA1 to
these cells produced selective changes in specific markers of cell activation (e.g., [Ca2+]i and L-
selectin shedding), and these appeared to be genuinely linked to the process of PMN apoptosis,
which is spontaneously switched on following in vitro cell incubations. Finally, all the effects
observed here are likely to be due to the specific N terminus region of the ANXA1 protein, as
evinced by comparing the actions of the native protein to those of human recombinant ANXA5
and the chimeric protein ANXA1-5.
For many years, inhibition of phospholipase A2 was believed to be the principal mechanism for
the anti-inflammatory properties of ANXA1 (6, 13). This mechanism of action was subsequently
challenged by reports showing that ANXA1 and ANXA1-derived peptides inhibited PMN
recruitment in animal models in which inhibitors of arachidonic acid catabolism were inactive
(10, 45). Subsequent studies demonstrated that these pharmacological antimigratory actions also
had pathophysiological relevance, because the endogenous protein displayed similar
antimigratory effects (16, 46). The model proposed for the PMN was that upon cell adhesion on
an endothelium, the protein would be exported from the cytosol onto the cell surface and at this

level it would operate through an unknown mechanism to down-regulate the extent of PMN
emigration across the endothelium (47). In an in vivo context, exogenous and endogenous
ANXA1 provoked detachment of adherent PMN (11, 46).
More recently, two membrane targets for ANXA1 have been identified: One is the fMLP
receptor or FPR (19), and the other is the α4β1 integrin (17). It is possible that both molecular
targets could be responsible for the antimigratory actions of ANXA1 discussed previously,
although it is more likely that the interaction with α4β1 integrin is more relevant to cells of the
mono-myelocytic lineage (17, 48). In addition, more recent data indicate that FPR might not be
the sole receptor for ANXA1 and its peptides (22). The present study was planned to investigate
ANXA1 effects that would be produced downstream from the interaction with its cell surface
target(s). In this manner, we would fill a gap in the field of ANXA1 biology and PMN functions.
Historically, several members of the annexin family (ANXA1, 2, 5–7, and 12) have been
reported to possess Ca2+ channel activities (49, 50). These activities were investigated only in in
vitro settings and often with reconstituted systems and were possibly mediated by the large core
common to these proteins. The core retains the highest homology among members of the
annexin superfamily, and it is thought to be responsible for annexin ability to bind phospholipids
in the presence of Ca2+ (1). In fact, the comparison of Ca2+ transport activities of several
annexins was conducted using large unilamellar vesicles in which purified annexins were
inserted. In these experiments, ANXA5 was reported to be a much better transporter than
ANXA1 (threefold difference at any tested concentration) (51). In a more physiological context,
ANXA5 was also shown to mediate peroxide-induced Ca2+ influx in B lymphocytes (52).
A recent study by Walther et al. (19) used ANXA1-derived peptides to mobilize Ca2+ in human
PMN, using concentrations in the 10–20 µM range. The present study began by testing the effect
of submicromolar concentrations of full-length ANXA1 added to human PMN in a test tube.
ANXA1 produced transient changes in [Ca2+]i; this effect was concentration-dependent and
required an intact ANXA1 N terminus, as demonstrated by the results produced with human
recombinant ANXA5 and the chimeric protein ANXA1-5. ANXA1-induced [Ca2+]i flux was
linked to a modest chemotactic effect and, more clearly, to shedding of L-selectin from the cell
surface. The ANXA1 concentrations used in these in vitro experiments are lower than the
maximal extracellular concentration reported so far for the protein (2 µM) (53). In addition,
resting PMN have high intracellular levels of ANXA1 (∼3–4% of total cytosolic protein) (5), and
the large majority of this (∼50–70%) is exported on the plasma membrane upon cell adhesion
(16). Thus, it is likely that high local concentrations of the protein can be reached on the PMN
cell surface within the microenvironment of an adherent PMN.
In view of the recent data linking ANXA1 to receptors of the FPR family, in most of the
experimental settings we used fMLP as an internal control. Data obtained in the experiments of
[Ca2+]i flux and second-messenger generation indicated a disparity of effects between the two
PMN activators. Note that at variance from Walther et al. (19), Strausbaugh and Rosen (18)
failed to demonstrate that ANXA1-induced L-selectin shedding was brought about by an
interaction with FPR. This is in contrast to the data obtained in the present study, in which Boc1
blocked fMLP and ANXA1-induced L-selectin shedding. The reason for this discrepancy may
lie in the enantiomer used, because only the D-Leu containing Boc derivatives act as antagonists
at this receptor (54) as well as the lipoxin A4 receptor (22). ANXA1-induced [Ca2+]i flux was

blocked by the FPR antagonist Boc1 and the calcium-entry blocker SKF-96365. The same
applied to ANXA1-induced L-selectin shedding, indicating a mechanistic link to receptor
activation and transient changes in intracellular calcium.
PMN are terminally differentiated cells with a short life span once released from the bone
marrow in the blood stream. In a matter of hours, they undergo programmed cell death and are
then rapidly removed from circulation by the reticulo-endoplasmic system. PMN apoptosis may
be either accelerated or delayed by multiple specific agents (reviewed in ref 55), and delayed
PMN apoptosis is associated with several inflammatory diseases characterized by sustained
neutrophil activation (56). Interestingly, a functional association between UV-induced PMN
apoptosis and activation of the metalloprotease that mediates shedding of L-selectin has been
proposed (57). This is possibly associated, at least in part, to the fact that engagement of PMN
adhesion molecules had already been shown to modulate programmed cell death (58). For
instance, cross-linking of CD11a or CD11b, but not CD18, with specific mAbs produced a delay
in spontaneous PMN apoptosis (59). In contrast, cross-linking of L-selectin with its sulfatide
ligands accelerated the degree of the apoptotic process (59). For these reasons, we tested the
effect of exogenous ANXA1 on PMN apoptosis, reporting, for the first time, a promoting effect
of the protein.
ANXA1-induced acceleration of PMN apoptosis occurred in a concentration-dependent manner,
and it was demonstrated using three distinct detection systems. The concentrations required were
in line with those necessary to activate the cell in terms of [Ca2+]i flux and L-selectin shedding.
Conversely, inhibitors of [Ca2+]i flux prevented ANXA1-induced apoptosis. Many cell types
have been shown to enter into apoptosis following a rise in intracellular Ca2+, for example,
thymocytes and eosinophils (55). However, sustained elevation of [Ca2+]i with calcium
ionophores has also been associated with a retardation in neutrophil apoptosis (60). It is plausible
that different kinetics, amplitudes, or compartmentalization of the Ca2+ fluxes may generate
opposite effects on PMN apoptosis. Alternatively, the proapoptotic effect of ANXA1 seen here
may not depend on the intracellular Ca2+ rise per se but on another yet unidentified ANXA1-
promoted signal, which is Ca2+-dependent.
This proapoptotic effect of ANXA1 resembles more that produced by TNF-β rather than fMLP.
fMLP induces the process of apoptosis in a superoxide release-dependent pathway, whereas
TNF-β-induced apoptosis appears to be unrelated to activation of the respiratory burst oxidase
(61). ANXA1-induced PMN apoptosis was linked to Bad dephosphorylation. It is thought that
following an increase in cytosolic Ca2+, the Ca2+-activated protein phosphatase calcineurin
dephoshorylates Bad, which allows Bad to associate with mitochondria, heterodimerize with
Bcl-XL, and promote apoptosis (62, 63). The endpoint would be accelerated apoptosis. Our data
are in accord with this finding, and further studies will clarify this pathway in more detail.
The following scheme summarizes the findings reported here in the context of PMN migration in
inflammation. Figure 10 shows the current working hypothesis. Several studies have shown that
upon PMN adhesion ANXA1 moves to the cell surface compartment, and here it acts, probably
through FPR or related receptors, to cause cell detachment, thereby reducing the extent of PMN
extravazation. This has been demonstrated for both exogenous (11) and endogenous (46)
ANXA1. This mechanism of action may well be responsible for the role that ANXA1 plays in
the phase of resolution of inflammation. Accordingly, passive immunization of animals against

ANXA1 prolongs the extent of PMN recruitment into a specific tissue site (12). Here, we report
that besides promoting controlled PMN activation (likely to be necessary to induce the cell
detachment phenomenon as discussed previously), ANXA1 accelerates the induction of PMN
apoptosis. This may impact on PMN biology and resolution of inflammation in two manners.
First, ANXA1-detached PMN will enter into apoptosis and clear from the blood stream in a
much faster and efficient manner. It is likely that previously adherent PMN may somehow be
primed and at least in part activated, hence these cells could be of some danger; their quick
removal by the nonphlogistic process of apoptosis (Fig. 10, option 1) is therefore more
compatible with host survival. Second, ANXA1 may promote apoptosis of PMN that are already
at the site of inflammation (Fig. 10, option 2). We now know that once extravazated, PMN
switch on ANXA1 synthesis (15), and that higher intracellular ANXA1 levels have been
associated with a degree of susceptibility to apoptosis (34). In addition, and more relevant to the
data here, produced with exogenously applied ANXA1, extracellular ANXA1 is present at the
site of inflammation, and its expression is associated with PMN influx (15, 64–66). More recent
kinetics studies indicate the presence of higher ANXA1 levels in inflammatory exudates in
concomitance with the phase of resolution, that is, the lowering of PMN numbers after the peak
of cell influx (22).
Glucocorticoids are known to retard neutrophil apoptosis (67, 68). As ANXA1 is a
glucocorticoid-regulated protein, it is difficult to reconciliate these findings with the data
presented here. It is tempting to propose that glucocorticoid regulation of PMN ANXA1 may be
more relevant in the context of cell adhesion and detachment (46), whereas exudates expression
of the protein could be mostly due to the cell exudation process, and less susceptible to
glucocorticoid regulation. For instance, extracellular ANXA1 can have different effects from the
intracellular protein. Studies with cells expressing different ANXA1 levels, and their potential
susceptibility to glucocorticoid-mediated regulation of apoptosis, will help clarify this apparent
contrasting point.
ACKNOWLEDGMENTS
We thank Dr. M. Tissot for the calcium analysis and Mrs. S. Wheller for technical support. This
work was sponsored by the CNRS and the Wellcome Program Grant 051887/Z/97/A. M.P. is a
senior research fellow of the Arthritis Research Campaign UK (grant P0583). We thank Dr. H.R.
Krell of Roche Diagnostic GmbH Pharma Research Penzberg Nonnenwald (Germany) for the
generous gift of compounds RO31-9790 and RO28-2653.
REFERENCES
1. Raynal, P., and Pollard, H. B. (1994) Annexins: the problem of assessing the biological role
for a gene family of multifunctional calcium- and phospholipid-binding proteins. Biochim.
Biophys. Acta 1197, 63–93
2. Gerke, V., and Moss, S. E. (2002) Annexins: from structure to function. Physiol. Rev. 82,
331–371
3. Donnelly, S. R., and Moss, S. E. (1997) Annexins in the secretory pathway. Cell. Mol. Life
Sci. 53, 533–538

4. Goulding, N. J., Luying, P., and Guyre, P. M. (1990) Characteristics of lipocortin 1 binding
to the surface of human peripheral blood leucocytes. Biochem. Soc. Trans. 18, 1237–1238
5. Francis, J. W., Balazovich, K. J., Smolen, J. E., Margolis, D. I., and Boxer, L. A. (1992)
Human neutrophil annexin I promotes granule aggregation and modulates Ca(2+)-dependent
membrane fusion. J. Clin. Invest. 90, 537–544
6. Flower, R. J. (1988) Lipocortin and the mechanism of action of the glucocorticoids. Br. J.
Pharmacol. 94, 987–1015
7. Solito, E., Raugei, G., Melli, M., and Parente, L. (1991) Dexamethasone induces the
expression of the mRNA of lipocortin 1 and 2 and the release of lipocortin 1 and 5 in
differentiated, but not undifferentiated U-937 cells. FEBS Lett. 291, 238–244
8. Peers, S. H., and Flower, R. J. (1991) Site of anti-inflammatory action of dexamethasone in
rabbit skin. Eur. J. Pharmacol. 196, 37–41
9. Goulding, N. J., Godolphin, J. L., Sharland, P. R., Peers, S. H., Sampson, M., Maddison, P.
J., and Flower, R. J. (1990) Anti-inflammatory lipocortin 1 production by peripheral blood
leucocytes in response to hydrocortisone. Lancet 335, 1416–1418
10. Perretti, M., Ahluwalia, A., Harris, J. G., Goulding, N. J., and Flower, R. J. (1993)
Lipocortin-1 fragments inhibit neutrophil accumulation and neutrophil-dependent edema in
the mouse. A qualitative comparison with an anti-CD11b monoclonal antibody. J. Immunol.
151, 4306–4314
11. Lim, L. H., Solito, E., Russo-Marie, F., Flower, R. J., and Perretti, M. (1998) Promoting
detachment of neutrophils adherent to murine postcapillary venules to control inflammation:
effect of lipocortin 1. Proc. Natl. Acad. Sci. USA 95, 14535–14539
12. Perretti, M., Ahluwalia, A., Harris, J. G., Harris, H. J., Wheller, S. K., and Flower, R. J.
(1996) Acute inflammatory response in the mouse: exacerbation by immunoneutralization of
lipocortin 1. Br. J. Pharmacol. 117, 1145–1154
13. Russo-Marie, F. (1992) Annexins, phospholipase A2 and the glucocorticoids. In Annexins,
Portland Press, London
14. Perretti, M., Christian, H., Wheller, S. K., Aiello, I., Mugridge, K. G., Morris, J. F., Flower,
R. J., and Goulding, N. J. (2000) Annexin I is stored within gelatinase granules of human
neutrophil and mobilized on the cell surface upon adhesion but not phagocytosis. Cell Biol.
Int. 24, 163–174
15. Oliani, S. M., Paul-Clark, M. J., Christian, H. C., Flower, R. J., and Perretti, M. (2001)
Neutrophil interaction with inflamed postcapillary venule endothelium alters annexin 1
expression. Am. J. Pathol. 158, 603–615

16. Perretti, M., Croxtall, J. D., Wheller, S. K., Goulding, N. J., Hannon, R., and Flower, R. J.
(1996) Mobilizing lipocortin 1 in adherent human leukocytes downregulates their
transmigration. Nat. Med. 2, 1259–1262
17. Solito, E., Romero, I. A., Marullo, S., Russo-Marie, F., and Weksler, B. B. (2000) Annexin1
binds to U937 monocytic cells and inhibits their adhesion to microvascular endothelium::
Involvement of the α4β1 integrin. J. Immunol. 165, 1573–1581
18. Strausbaugh, H. J., and Rosen, S. D. (2001) A potential role for annexin 1 as a physiologic
mediator of glucocorticoid-induced L-selectin shedding from myeloid cells. J. Immunol.
166, 6294–6300
19. Walther, A., Riehemann, K., and Gerke, V. (2000) A novel ligand of the formyl peptide
receptor: annexin I regulates neutrophil extravasation by interacting with the FPR. Mol. Cell
5, 831–840
20. Dalpiaz, A., Ferretti, M. E., Vertuani, G., Traniello, S., Scatturin, A., and Spisani, S. (2002)
C- and N-terminal residue effect on peptide derivatives' antagonism toward the formyl-
peptide receptor. Eur. J. Pharmacol. 436, 187–196
21. Le, Y., Oppenheim, J. J., and Wang, J. M. (2001) Pleiotropic roles of formyl peptide
receptors. Cytokine Growth Factor Rev. 12, 91–105
22. Perretti, M., Chiang, N., La, M., Fierro, I. M., Marullo, S., Getting, S. J., Solito, E., and
Serhan, C. N. (2002) Endogenous lipid- and peptide-derived anti-inflammatory pathways
generated with glucocorticoid and aspirin treatment activate the lipoxin A(4) receptor. Nat.
Med. 8, 1296–1302
23. Perretti, M., Wheller, S. K., Choudhury, Q., Croxtall, J. D., and Flower, R. J. (1995)
Selective inhibition of neutrophil function by a peptide derived from lipocortin 1 N-
terminus. Biochem. Pharmacol. 50, 1037–1042
24. Zouki, C., Ouellet, S., and Filep, J. G. (2000) The anti-inflammatory peptides, antiflammins,
regulate the expression of adhesion molecules on human leukocytes and prevent neutrophil
adhesion to endothelial cells. FASEB J. 14, 572–580
25. Platt, N., da Silva, R. P., and Gordon, S. (1998) Recognizing death: the phagocytosis of
apoptotic cells. Trends Cell Biol. 8, 365–372
26. Boyum, A. (1968) A one-stage procedure for isolation of granulocytes and lymphocytes
from human blood. General sedimentation properties of white blood cells in a 1g gravity
field. Scand. J. Clin. Lab. Invest. Suppl. 97, 51–76
27. Pozzan, T., Lew, D. P., Wollheim, C. B., and Tsien, R. Y. (1983) Is cytosolic ionized
calcium regulating neutrophil activation? Science 221, 1413–1415
28. Grynkiewicz, G., Poenie, M., and Tsien, R. Y. (1985) A new generation of Ca2+ indicators
with greatly improved fluorescence properties. J. Biol. Chem. 260, 3440–3450

29. Tissot, M., Mathieu, J., Mirossay, L., Thuret, A., and Giroud, J. P. (1991)
Polyphosphoinositide metabolism in polymorphonuclear cells from healthy and thermally
injured rats: effect of the immunomodulator RU 41740. J. Leukoc. Biol. 50, 607–614
30. Berridge, M. J., Dawson, R. M., Downes, C. P., Heslop, J. P., and Irvine, R. F. (1983)
Changes in the levels of inositol phosphates after agonist-dependent hydrolysis of membrane
phosphoinositides. Biochem. J. 212, 473–482
31. Johnston, R. B., Jr., Keele, B. B., Jr., Misra, H. P., Lehmeyer, J. E., Webb, L. S., Baehner,
R. L., and RaJagopalan, K. V. (1975) The role of superoxide anion generation in phagocytic
bactericidal activity. Studies with normal and chronic granulomatous disease leukocytes. J.
Clin. Invest. 55, 1357–1372
32. Keller, H. U., Gerber, H., Hess, M. W., and Cottier, H. (1976) Studies on the regulation of
the neutrophil chemotactic response using a rapid and reliable method for measuring random
migration and chemotaxis of neutrophil granulocytes. Agents Actions 6, 326–339
33. Zigmond, S. H., and Hirsch, J. G. (1973) Leukocyte locomotion and chemotaxis. New
methods for evaluation, and demonstration of a cell-derived chemotactic factor. J. Exp. Med.
137, 387–410
34. Solito, E., de Coupade, C., Canaider, S., Goulding, N. J., and Perretti, M. (2001)
Transfection of annexin 1 in monocytic cells produces a high degree of spontaneous and
stimulated apoptosis associated with caspase-3 activation. Br. J. Pharmacol. 133, 217–228
35. Elstein, K. H., and Zucker, R. M. (1994) Comparison of cellular and nuclear flow cytometric
techniques for discriminating apoptotic subpopulations. Exp. Cell Res. 211, 322–331
36. Huber, R., Romisch, J., and Paques, E. (1990) The crystal and molecular structure of human
annexin V, an anticougulant protein that binds to calcium and membrane. EMBO J. 9, 3867–
3974
37. Huber, R., Berendes, R., Burger, A., Schneider, M., Karshikov, A., Luecke, H., Romisch, J.,
and Paques, E. (1992) Crystal and molecular structure of human annexin V after refinement.
Implications for structure, membrane binding and ion channel formation of the annexin
family of proteins. J. Mol. Biol. 223, 683–704
38. Andersson, T., Dahlgren, C., Pozzan, T., Stendahl, O., and Lew, P. D. (1986)
Characterization of fMet-Leu-Phe receptor-mediated Ca2+ influx across the plasma
membrane of human neutrophils. Mol. Pharmacol. 30, 437–443
39. Lew, D. P. (1989) Receptor signalling and intracellular calcium in neutrophil activation.
Eur. J. Clin. Invest. 19, 338–346
40. Preece, G., Murphy, G., and Ager, A. (1996) Metalloproteinase-mediated regulation of L-
selectin levels on leucocytes. J. Biol. Chem. 271, 11634–11640

41. Peschon, J. J., Slack, J. L., Reddy, P., Stocking, K. L., Sunnarborg, S. W., Lee, D. C.,
Russell, W. E., Castner, B. J., Johnson, R. S., Fitzner, J. N., et al. (1998) An essential role
for ectodomain shedding in mammalian development. Science 282, 1281–1284
42. Fadeel, B., Ahlin, A., Henter, J. I., Orrenius, S., and Hampton, M. B. (1998) Involvement of
caspases in neutrophil apoptosis: regulation by reactive oxygen species. Blood 92, 4808–
4818
43. Maianski, N. A., Roos, D., and Kuijpers, T. W. (2003) Tumor necrosis factor alpha induces
a caspase-independent death pathway in human neutrophils. Blood 101, 1987–1995
44. Akgul, C., Moulding, D. A., and Edwards, S. W. (2001) Molecular control of neutrophil
apoptosis. FEBS Lett. 487, 318–322
45. Perretti, M., and Flower, R. J. (1993) Modulation of IL-1-induced neutrophil migration by
dexamethasone and lipocortin 1. J. Immunol. 150, 992–999
46. Mancuso, F., Flower, R. J., and Perretti, M. (1995) Leukocyte transmigration, but not rolling
or adhesion, is selectively inhibited by dexamethasone in the hamster post-capillary venule.
Involvement of endogenous lipocortin 1. J. Immunol. 155, 377–386
47. Perretti, M. (1997) Endogenous mediators that inhibit the leukocyte-endothelium
interaction. Trends Pharmacol. Sci. 18, 418–425
48. Perretti, M., Ingegnoli, F., Wheller, S. K., Blades, M. C., Solito, E., and Pitzalis, C. (2002)
Annexin 1 modulates monocyte-endothelial cell interaction in vitro and cell migration in
vivo in the human SCID mouse transplantation model. J. Immunol. 169, 2085–2092
49. Pollard, H., Rojas, E., Merezhinskaya, N., Kuijpers, G., Srivastava, M., Zhang-Keck, Z.,
Shirvan, A., and Burns, A. (1992) Synexin (annexin VII), Portland Press Research, London
50. Luecke, H., Chang, B. T., Mailliard, W. S., Schlaepfer, D. D., and Haigler, H. T. (1995)
Crystal structure of the annexin XII hexamer and implications for bilayer insertion. Nature
378, 512–515 (see comments)
51. Matsuda, R., Kaneko, N., and Horikawa, Y. (1997) Presence and comparison of Ca2+
transport activity of annexins I, II, V, and VI in large unilamellar vesicles. Biochem.
Biophys. Res. Commun. 237, 499–503
52. Kubista, H., Hawkins, T. E., Patel, D. R., Haigler, H. T., and Moss, S. E. (1999) Annexin 5
mediates a peroxide-induced Ca(2+) influx in B cells. Curr. Biol. 9, 1403–1406
53. Christmas, P., Callaway, J., Fallon, J., Jones, J., and Haigler, H. T. (1991) Selective
secretion of annexin 1, a protein without a signal sequence, by the human prostate gland. J.
Biol. Chem. 266, 2499–2507
54. Dalpiaz, A., Ferretti, M. E., Pecoraro, R., Fabbri, E., Traniello, S., Scatturin, A., and Spisani,
S. (1999) Phe-D-Leu-Phe-D-Leu-Phe derivatives as formylpeptide receptor antagonists in

human neutrophils: cellular and conformational aspects. Biochim. Biophys. Acta 1432, 27–
39
55. Ward, I., Dransfield, I., Chilvers, E. R., Haslett, I., and Rossi, A. G. (1999) Pharmacological
manipulation of granulocyte apoptosis: potential therapeutic targets. Trends Pharmacol. Sci.
20, 503–509
56. Dibbert, B., Weber, M., Nikolaizik, W. H., Vogt, P., Schoni, M. H., Blaser, K., and Simon,
H. U. (1999) Cytokine-mediated Bax deficiency and consequent delayed neutrophil
apoptosis: a general mechanism to accumulate effector cells in inflammation. Proc. Natl.
Acad. Sci. USA 96, 13330–13335
57. Kern, P. M., Keilholz, L., Forster, C., Stach, C., Beyer, T. D., Gaipl, U. S., Kalden, J. R.,
Hallmann, R., and Herrmann, M. (2000) UVB-irradiated T-cells undergoing apoptosis lose
L-selectin by metalloprotese-mediated shedding. Int. J. Radiat. Biol. 76, 1265–1271
58. Dransfield, I., Stocks, S. C., and Haslett, C. (1995) Regulation of cell adhesion molecule
expression and function associated with neutrophil apoptosis. Blood 85, 3264–3273
59. Watson, R. W., Rotstein, O. D., Nathens, A. B., Parodo, J., and Marshall, J. C. (1997)
Neutrophil apoptosis is modulated by endothelial transmigration and adhesion molecule
engagement. J. Immunol. 158, 945–953
60. Whyte, M. K., Hardwick, S. J., Meagher, L. C., Savill, J. S., and Haslett, C. (1993) Transient
elevations of cytosolic free calcium retard subsequent apoptosis in neutrophils in vitro. J.
Clin. Invest. 92, 446–455
61. Kettritz, R., Falk, R. J., Jennette, J. C., and Gaido, M. L. (1997) Neutrophil superoxide
release is required for spontaneous and FMLP- mediated but not for TNF alpha-mediated
apoptosis. J. Am. Soc. Nephrol. 8, 1091–1100
62. Wang, H. G., Pathan, N., Ethell, I. M., Krajewski, S., Yamaguchi, Y., Shibasaki, F.,
McKeon, F., Bobo, T., Franke, T. F., and Reed, J. C. (1999) Ca2+-induced apoptosis
through calcineurin dephosphorylation of BAD. Science 284, 339–343
63. Porter, A. G. (1999) Protein translocation in apoptosis. Trends Cell Biol. 9, 394–401 (In
Process Citation)
64. Tsao, F. H., Meyer, K. C., Chen, X., Rosenthal, N. S., and Hu, J. (1998) Degradation of
annexin I in bronchoalveolar lavage fluid from patients with cystic fibrosis. Am. J. Respir.
Cell Mol. Biol. 18, 120–128
65. Vergnolle, N., Comera, C., and Bueno, L. (1995) Annexin 1 is overexpressed and
specifically secreted during experimentally induced colitis in rats. Eur. J. Biochem. 232,
603–610

66. Vergnolle, N., Comera, C., More, J., Alvinerie, M., and Bueno, L. (1997) Expression and
secretion of lipocortin 1 in gut inflammation are not regulated by pituitary-adrenal axis. Am.
J. Physiol. 273, R623–R629
67. Meagher, L. C., Cousin, J. M., Seckl, J. R., and Haslett, C. (1996) Opposing effects of
glucocorticoids on the rate of apoptosis in neutrophilic and eosinophilic granulocytes. J.
Immunol. 156, 4422–4428
68. Cox, G., and Austin, R. C. (1997) Dexamethasone-induced suppression of apoptosis in
human neutrophils requires continuous stimulation of new protein synthesis. J. Leukoc. Biol.
61, 224–230
Received December 12, 2002; accepted May 2, 2003.

Fig. 1
Figure 1. ANXA1 activates transient [Ca2+]i fluxes in human PMN. A) PMN were loaded with Fura-2 and stimulated
with the indicated compounds, and the resulting [Ca2+]i-dependent fluorescence changes were recorded. Shown is a
comparison of the Ca2+ signals promoted by 10 nM fMLP and 500 nM ANXA1. B) Concentration-related response for
exogenous ANXA1 on [Ca2+]i. C) Comparison of the Ca2+ signals promoted by ANXA1, annexin 5 (ANXA5), and chimera
ANXA1-5 (all tested at 500 nM). D) Signals recorded in the presence of 50 µM SKF-96365 (SKF). E) Concentration-
related response for the Ca2+ influx blocker SKF-96365 on the ANXA1 response. A, D) Recordings are from single donor
samples and are representative of results obtained from 3 to 10 different donors. B, E) *P<0.05 vs. unstimulated cells (n=3
experiments).

Fig. 2
Figure 2. ANXA1-induced [Ca2+]i in PMN is blocked by Boc1 and SKF-96365. A, B) PMN were loaded with Fura-2
and [Ca2+]i signals promoted by fMLP (100 nM) or ANXA1 (500 nM). Recordings in the absence or presence of Boc1 (20
µM) are shown. C) Cross-desensitization between fMLP (100 nM) and ANXA1 (500 nM). In all cases, experiments
correspond to recordings on single donor samples and are representative of results obtained from ≥3 different donors.

Fig. 3
Figure 3. Comparison of PMN responses elicited by fMLP and ANXA1. A) PMN were labeled with [3H]-myo-
inositol and incubated at 37°C with 100 nM fMLP or 500 nM ANXA1 for various times (30 s and 1 min for fMLP; 30 s
and 1, 5, and 10 min for ANXA1). In some experiments, PMN were preincubated with ANXA1 for 10 min before
addition of fMLP. Inositol phosphate analysis was performed as indicated in Materials and Methods. Results,
corresponding to the maximal observed stimulation (30 s for fMLP, 5 min for ANXA1), are the mean ±SE of triplicates in
one experiment and are representative of three independent experiments; *P<0.05 vs. control inositol species generation,
as calculated on original data. B) Superoxide anion production by PMN incubated for 10 min at 37°C with 100 nM fMLP
and/or the indicated concentrations of ANXA1, as measured by the reduction of ferricytochrome C. Histograms are the
mean of triplicates and are representative of three independent experiments. C) PMN chemotaxis promoted by fMLP and
ANXA1. Cells were added to the upper compartment of Boyden chambers with 10 nM fMLP alone or with the indicated
concentrations of ANXA1 being added to the lower compartment. Chambers were incubated for 90 min at 37°C. Cell
migration was determined by measuring the distance of the leading front (see Materials and Methods), and specific
migration was calculated by subtracting the distance of the leading front measured in control chambers in the absence of
chemoattractant. Data are the mean ±SE of three independent experiments performed in triplicate. ANOVA test, *P<0.05
vs. unstimulated cells.

Fig. 4
Figure 4. ANXA1 induces L-selectin shedding from human PMN cell surface. A) Single donor FACS profile
comparing ANXA1, ANXA5, and the chimera ANXA1-5 (0.7 µM in all cases), with the response produced by 10 nM
fMLP on L-selectin PMN surface expression. B) Cumulative data performed with three different donors in triplicate.
ANOVA test, *P<0.05 vs. relative basal L-selectin expression (B group).

Fig. 5
Figure 5. Effect of SKF-96365, Boc1, and RO31-9790 on ANXA1-induced L-selectin shedding. Whole blood
aliquots were pretreated with SKF-96365 (25 µM; A), Boc1 (20 µM; B), RO31-9790 or RO28-2653 (50 µM; C), or
vehicle for 3 min followed by fMLP (10 nM) or ANXA1 (0.7 µM) for 30 min at 37°C. PMN surface L-selectin levels
were measured by flow cytometry. Data represent the means ± SE of three experiments (three different donors) performed
in triplicate. ANOVA test, #P< 0.05 vs. corresponding vehicle group, *P<0.05 vs. relative basal L-selectin expression
(85±15 MFI units, C).

Fig. 6
Figure 6. ANXA1 induces apoptosis in human neutrophils. Human PMN were incubated with or without 500 nM
ANXA1 for 6 h at 37°C. The degree of cell apoptosis was measured in three distinct manners. A) Representative FACS
analysis of control and ANXA1-treated PMN obtained with PI DNA staining. B) Representative Hoechst immunostaining
of the previous donor. C) Dot plots showing ANXA5 (FL1 channel) and PI (FL2 channel) staining. Data shown in the
three panels have been repeated at least with 8–10 different donors.

Fig. 7
Figure 7. Comparative effect on neutrophils apoptosis induced by ANXA1, ANXA5, and chimeric ANXA1-5. A)
Effect of 1 nM PMA, 10 nM fMLP, or 500 nM ANXA1 on PMN apoptosis followed by 6 h incubation as determined with
the PI cell cycle technique. B) As in A, but after cell incubation with 500 nM of each recombinant annexin. ANOVA test,
*P<0.05 from untreated PMN (basal). #P<0.05 from ANXA1-treated PMN. Data are mean ±SE of five experiments.

Fig. 8
Figure 8. Calcium blockers inhibit ANXA1-induced apoptosis in PMN. Human PMN were incubated with or without
ANXA1 and the reported inhibitors/antagonists for 6 h at 37°C. A) Inhibition of ANXA1 concentration-response curve for
PMN apoptosis by the Ca2+ influx blocker SKF-96365 (SKF; 25 µM). B) ANXA1-promoted apoptosis was reverted by
the Ca2+ calcium chelator BAPTA (3 µM) but not by the L-selectin shedding inhibitors RO31-9790 (RO; 50 µM). PMN
apoptosis was determined with ANXA5/PI staining; ANOVA test, *P<0.005 from ANXA1-treated PMN. Data are mean
±SE of three experiments.

Fig. 9
Figure 9. ANXA1 induces BAD dephosphorylation. Detection of BAD and phospho-BAD in human neutrophils
following a 30-min treatment with human recombinant ANXA1 (500 nM) in the presence or absence of SKF-96365 (25
µM). A) Upper panel: phospho-BAD immunoreactivity as detected with an antibody to phospho-serine; lower panel: total
BAD immunoreactivity. Western blot profiles are from a single donor. C) Histogram (% of phosphorylated BAD) was
obtained as mean ±SE from three donors. ANOVA test, P<0.05 vs. untreated PMN.

Fig. 10
Figure 10. Hypothesis for ANXA1-induced PMN apoptosis in the context of an inflammatory scenario. There are
at least two options by which ANXA1 can affect PMN during inflammation. First, ANXA1 can cause detachment of
adherent PMN, likely through a process of controlled activation. It is possible that the “detached” portion of PMN will
enter into apoptosis and clear from the blood stream in a much faster and efficient manner. As adherent cells are also
activated, this mechanism is compatible with host survival. Second, ANXA1 may promote apoptosis of PMN that are
already at the site of inflammation, expressed as a dotted oval area. Many studies have shown ANXA1 expression in
inflammatory exudates; therefore, it is possible that the proapoptotic effect of the protein may be relevant to shortening
the life expectancy of extravasated cells (which otherwise is much longer in respect to PMN circulating in the blood). In
either case, ANXA1-induced PMN apoptosis will promote resolution of inflammation.