| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1-acid glycoprotein (AGP) induces rises in cytosolic Ca2+ in neutrophil granulocytes via sialic acid binding immunoglobulin-like lectins (Siglecs)
* Division of Pharmacology, Department of Medicine and Care, and
Division of Cell Biology, Department of Biomedicine and Surgery, Linköping University, Linköping, Sweden
1Correspondence: Division of Pharmacology, Department of Medicine and Care, Linköping University, S-581 85 Linköping, Sweden. E-mail: petgu{at}imv.liu.se
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1-acid glycoprotein (AGP) induces rises in [Ca2+]i in neutrophils and sought to identify the corresponding AGP receptor (or receptors). We found that AGP elicited a minimal rise in [Ca2+]i in Fura-2-loaded neutrophils, and this response was markedly enhanced by pretreatment with anti-L-selectin antibodies. (The EC50 value of the AGP-induced Ca2+ response was 9 µg/ml.) Activation of phospholipase-C, Src tyrosine kinases, and PI3 kinases proved to be essential for the AGP-mediated increase in [Ca2+]i, whereas the p38 MAPK and SYK signaling pathways were not involved. Furthermore, antibodies against sialic acid binding, immunoglobulin-like lectin 5 (Siglec-5) and oligosaccharide 3'-sialyl-lactose both antagonized the AGP-induced response and caused an immediate increase in [Ca2+]i in anti-L-selectin-treated neutrophils, which indicates a signaling capacity of Siglec-5. We used modified forms of AGP (treated with mild periodate or neuraminidase) to establish the importance of sialic acid residues. The modified forms of AGP caused a much smaller rise in [Ca2+]i than did unaltered AGP. Affinity chromatography confirmed that unchanged AGP, but not neuraminidase-treated AGP, bound to Siglec-5. Our report provides the first evidence for a signaling capacity by AGP through a defined receptor. Pre-engagement of L-selectin significantly enhanced this signaling capacity.—Gunnarsson, P., Levander, L., Påhlsson, P., Grenegård, M. The acute-phase protein 1-acid glycoprotein (AGP) induces rises in cytosolic Ca2+ in neutrophil granulocytes via sialic acid binding immunoglobulin-like lectins (Siglecs).
Key Words: orosomucoid • plasma protein • calcium signaling • carbohydrate • L-selectin • phagocyte
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1-acid glycoprotein (AGP) or orosomucoid, is found in human plasma at a concentration of 
1 mg/ml, and several studies have indicated it is involved in modulating inflammation and immune responses. For instance, studies of cell function in vitro have shown that the effects of AGP on neutrophils include reduction of phagocytic capacity (1)
as well as inhibition of superoxide anion production and migration (2
, 3)
. AGP also enhances the ability of macrophages to inhibit the proliferative effects of interleukin-1 (IL-1) on murine thymocytes (4)
, reduces platelet aggregation (5)
, and impedes the anti-CD3-stimulated proliferative response in lymphocytes (6)
. However, the exact mechanisms by which AGP modulates leukocytes and other cell types associated with inflammation are unknown at this time.
AGP carries five N-linked, complex-type oligosaccharide glycans that account for 45% of the molecular weight of the molecule (7)
. The plasma concentration and carbohydrate composition of AGP are altered in several patho-physiological conditions. For example, AGP exhibits decreased branching of the attached N-glycans during acute inflammation (8)
whereas increased branching and an elevated degree of fucosylation are commonly seen in patients with chronic inflammatory diseases such as rheumatoid arthritis (9)
. Furthermore, some in vitro studies have indicated that the carbohydrate side chains of AGP are involved in modulating leukocyte responses (3
, 4
, 6)
. However, the exact connection between the carbohydrate composition of AGP and the mechanisms by which this glycoprotein affects leukocytes remain to be established.
Based on current knowledge of AGP and its actions on leukocytes, we hypothesized that the carbohydrate side chains of the molecule recognize lectins on the surface of neutrophils. The N-linked glycans on AGP display the selectin binding determinant sialyl Lewis x (SLex); thus, AGP may influence neutrophil responses by attaching to L-selectin exposed on the surface of the cells. This assumption is supported by a study that demonstrates an interaction between AGP and E-selectin (10)
. Sialic acid binding immunoglobulin-like lectins (Siglecs) represent another class of lectins, and it is believed that neutrophils express the subtypes Siglec-5 and Siglec-9 (11
, 12)
. AGP is highly sialylated, hence Siglecs may also serve as binding sites for AGP. Besides facilitating binding, L-selectin induces intracellular signaling, including increases in cytosolic Ca2+ concentration and protein tyrosine phosphorylation (e.g., MAPK) (13
14
15)
. The capacity of Siglec-5 and Siglec-9 in this regard is less well defined. However, Siglec-5 and Siglec-9 and other CD33-related Siglecs are known to contain protein tyrosine phosphatase recruiting, immunoreceptor tyrosine-based inhibitory motif (ITIM)-like structures, which suggests that they participate in cell signaling events (11
, 16)
.
In the present study we sought to determine whether AGP directly provokes intracellular signaling in neutrophils and to identify the corresponding surface receptor (or receptors). Particular emphasis was placed on a possible interaction between N-linked glycans on AGP and L-selectin and the Siglec family of cell surface lectins on neutrophils. Ca2+ is a ubiquitous intracellular messenger that plays crucial roles in regulating neutrophil functional responses; therefore, we focused our work on a plausible Ca2+-mobilizing action of AGP. The results revealed that AGP binds to Siglecs on neutrophils, which leads to an immediate rise in [Ca2+]i. Thus, AGP may be an endogenous ligand for Siglec-5 and a signaling molecule that participates directly in the regulation of neutrophil responses. Pre-engagement of L-selectin was central to significantly enhance the Ca2+ signaling induced by the AGP-Siglec-5 interaction.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Preparation of desialylated AGP and mild periodate-treated AGP
We performed desialylation of AGP by incubation with Clostridium perfringens neuraminidase (Sigma-Aldrich; 100 mU/mg protein) in 0.1 M acetic acid buffer (pH 5.0) for 16 h at 37°C. Desialylated AGP (desialAGP) was desalted by Millipore Centricon centrifugation (Millipore, Bedford, MA, USA). To determine protein concentrations, a bicinchoninic acid assay was used according to the manufacturer’s instructions (Pierce Biotechnology Inc., Rockford, IL, USA) and calculations based on a standard curve of AGP in the range 50–1000 µg/ml were performed. AGP with modified sialic acid residues was produced by treatment with mild periodate without affecting the charge of the molecule. Briefly, AGP (5 mg) was incubated with 2 mM NaIO4 dissolved in PBS (pH 7.5), then 10% (v/v) ethylene glycol and 20 mM NaBH4 dissolved in PBS were added. The samples were desalted on PD-10 columns (GE Healthcare Bio-Sciences, Little Chalfont, UK), then lyophilized and stored at –70°C. The protein was reconstituted in ultrapure water before use.
Sialic acid analysis
The sialic acids were released from desialAGP (0.05 mg) by mild hydrolysis in 2 M acetic acid at 80°C for 3 h, as described by Varki and Diaz (17)
. The amount of sialic acid released was determined by high pH anion exchange chromatography (Dionex Carbopack PA-100 column at 15°C) with pulsed amperometric detection (HPAEC-PAD, Dionex, Sunnyvale, CA, USA). N-Glycolylneuraminic acid (Sigma-Aldrich) was used as an internal standard at a fixed concentration in the range of 42.4–62.2 µg/ml.
Isolation of neutrophils
A modified version of a method already described (18)
was used to isolate human neutrophils. Heparinized peripheral blood from healthy blood donors was centrifuged through a density gradient of PolymorphprepTM (Axis Shield PoC AS, Oslo, Norway) at 480 g for 40 min at room temperature in accordance with the manufacturer’s instructions. Separated neutrophils were resuspended in PBS (pH 7.3) and centrifuged at 480 g for 10 min at room temperature. Neutrophils were loaded with Fura-2 by adding 4 µM Fura-2-acetoxymethylester (AM) (Sigma-Aldrich) to the cell suspension, then incubating for 30 min at 37°C with gentle shaking. Remaining erythrocytes were removed by brief hypotonic lysis performed twice in distilled water. The cell suspension was kept cold during centrifugation and the isolated neutrophils were placed on ice until use. After the isolation procedure, neutrophils were suspended in Krebs-Ringer glucose buffer (pH 7.3) made up of 120 mM NaCl, 4.9 mM KCl, 1.2 mM MgSO4, 1.7 mM KH2PO4, 8.3 mM Na2HPO4, 1 mM CaCl2, and 10 mM glucose. The cell concentration was adjusted to 2 x 106 cells/ml and kept on ice until used. When preparing a neutrophil lysate, the Fura-2 loading step was omitted and the cells were lysed by exposure to a buffer (pH 8) containing 40 mM Tris (tris(hydroxymethyl)aminomethane)-HCl, 150 mM NaCl, 1% Triton-X, and 5 mM EDTA (ethylenediaminetetraacetic acid) and supplemented with 2 mM phenylmethanesulfonyl fluoride, 4 µM leupeptin, and 3 µM pepstatin. The lysate was incubated on ice for 30 min, then centrifuged for 20 min at 14,000 g and 4°C. A BCA assay (Pierce Biotechnology Inc.) was used to analyze the supernatant for protein concentration; the supernatant was then stored at –70°C.
Culture of HL-60 cells
Human acute leukemia myeloid (HL-60) cells were obtained from ATCC (Manassas, VA, USA), then cultured and induced to undergo neutrophil-like differentiation as described by Sjogren et al. (19)
. Briefly, the HL-60 cells were grown in RPMI 1640 containing glutamax supplemented with penicillin-streptomycin (50 U/ml; 50 µg/ml) and heat-inactivated FCS (10%). Differentiation into neutrophil-like cells was induced by supplementing the growth medium with 1.3% dimethyl sulfoxide (DMSO) for 6 days. The differentiated cells were loaded with Fura-2 and treated in the same manner as isolated neutrophils (described earlier in the "Isolation of neutrophils" section).
Measurement of cytosolic Ca2+ concentrations
Fluorescence signals from 2 ml aliquots of suspensions of neutrophils or HL-60 cells (2x106 cells/ml) were detected at 37°C under constant stirring (300 rpm) using a Hitachi F-2000 spectrofluorometer (Hitachi Ltd., Tokyo, Japan) especially designed to measure intracellular Ca2+ concentration. Fluorescence emission was registered at 510 nm during simultaneous excitation at 340 nm and 380 nm. The intracellular Ca2+ concentration was calculated by applying the general equation described by Grynkiewicz (20)
: [Ca2+]i = Kd(R–Rmin)/(Rmax–R)(Fo/Fs). Maximum and minimum ratios were determined by first adding 0.1% Triton X-100, then 25 mM EGTA (ethyleneglycoltetraacetic acid). Cell suspensions were incubated for 3 min at 37°C before each recording. In some experiments the cells were pretreated with the anti-L-selectin antibody DREG-56 (1 µg/ml) or an IgG1 isotype-matched antibody (clone 107.3, 1 µg/ml) for 3 min prior to the addition of AGP (0.5 mg/ml, 12 µM). After exposure to AGP for 3 min, the cells were stimulated with fMLP (100 nM) as a control of cell response. In the experiments using kinase inhibitors (i.e., Wortmannin, LY 294002, Piceatannol, SU 6656, PP2), the cells were incubated for 3 min at 37°C before they were challenged with DREG-56. In some series of experiments, the anti-Siglec-5 antibody 3'SL or 6'SL was added 2 min after DREG-56, and AGP was added 2 min later. In control experiments, vehicle was added instead of reagent at the same time intervals.
Affinity chromatography
A preclearing column was prepared by immobilization of 3 mg of BSA (Sigma-Aldrich) on a 1 ml High-Trap NHS-activated column (GE Healthcare Bio-Sciences) according to the manufacturer’s instructions. AGP affinity columns were generated by immobilization of 5 mg of AGP or 5 mg of desialylated AGP (17)
on 1 ml High-Trap NHS-activated columns. The amount of protein immobilized to each column was calculated and determined to be 2.75 mg for AGP and 2.6 mg for desialylated AGP. To reduce nonspecific interactions, neutrophil lysate (1 mg of protein) diluted in starting buffer (50 mM tris-HCl (pH 7.5) supplemented with 50 mM CaCl2) was precleared by passing it through the BSA column. The eluted proteins were collected and applied to either the AGP or the desialylated AGP affinity column connected to an AKTAprimeTM system (GE Healthcare Bio-Sciences) equipped with a UV detector recording absorbance at 280 nm. Nonbound proteins were eluted with starting buffer. The column was washed with starting buffer until A280 was back to baseline. Bound proteins were eluted with a 50 mM Tris-HCl buffer (pH 7.5) supplemented with 2.5 mM EDTA, followed by a 100 mM glycine-HCl buffer (pH 2.2). PD-10 columns (GE Healthcare Bio-Sciences) were used to desalt collected A280 absorbing fractions; the collected fractions were then lyophilized and stored at –70°C.
SDS-PAGE and Western blot
A 4–20% polyacrylamide gradient gel (Bio-Rad, Hercules, CA, USA) was used to separate proteins under denaturing conditions; the proteins were then transferred to immobilon-P membranes (Millipore). The membranes were blocked with rinsing buffer (50 mM Tris, 150 mM NaCl, 0.1% Tween-20) containing 5% dry milk, then incubated with polyclonal antibodies directed against human Siglec-5 and Siglec-9 (R&D Systems). A biotin-conjugated rabbit-anti-goat antibody (DAKO) was used as a secondary antibody; in a third step, a streptavidin-biotinylated HRP complex (GE Healthcare Bio-Sciences) was added. The membranes were rinsed in rinsing buffer after each incubation. The blots were developed with Western blot Luminol Reagent (Santa Cruz Biotechnology, Santa Cruz, CA, USA).
Flow cytometry analysis of Siglec-5 expression
Siglec-5 mAb (R&D Systems) were labeled with a Zenon mouse IgG labeling kit according to the manufacturer’s instruction (Molecular Probes). Isolated human neutrophils were pretreated with DREG-56 (1 µg/ml, 3 min) at 37°C before staining with Zenon-labeled Siglec-5 mAb. The cells were kept on ice until analyzed on FACSCalibur (BD Biosciences). Untreated cells and cells stained with Zenon-labeled unspecific antibodies were analyzed in parallel.
Statistics
Statistical analysis was conducted with GraphPad Prism 4 software (GraphPad Software Inc., San Diego, CA, USA). Significant differences were calculated by applying ANOVA and Bonferroni’s multiple comparison procedure as a post hoc test (*P<0.05, **P<0.01, and ***P<0.001). Results are presented as means ± SE. The EC50 value was estimated from the dose-response curve by nonlinear regression analysis of the data and is shown with the 95% confidence limit.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
shows that the addition of a physiological dose of AGP (0.5 mg/ml) induced only a minimal rise in [Ca2+]i (Fig. 1A
, trace 1). In fact, 15% (8 of 53) of the neutrophil preparations showed no response at all to such treatment (Fig. 1B
). Moreover, higher doses of AGP did not amplify the minor elevation of [Ca2+]i and lower doses did not cause any detectable Ca2+ response. Many low-efficacy activators are known to prime neutrophils, which renders them more responsive to other stimuli, such as chemotactic peptides. However, we found that exposure to AGP had no effect on a subsequent peak rise in [Ca2+]i induced by treatment with 100 nM fMLP (Fig. 1A
).
|
Engagement of L-selectin makes neutrophils and differentiated HL-60 cells sensitive to AGP
The N-linked glycans on AGP contain SLex epitopes, which are low-affinity ligands for L-selectin, and L-selectin has been shown to induce intracellular signaling in neutrophils (13
, 21
, 22)
. Accordingly, we hypothesized that binding of AGP to L-selectin may explain the minor increase in [Ca2+]i caused by AGP. To test that assumption, we pretreated neutrophils with a monoclonal antibody against L-selectin (DREG-56; 1 or 10 µg/ml), then exposed them to AGP (0.5 mg/ml). The Ca2+ response to AGP was markedly enhanced by DREG-56 pretreatment (original traces shown in Fig. 1A
, trace 2; results are summarized in Fig. 1B
). The AGP-provoked Ca2+ response was not influenced by addition of isotype-matched antibodies. Consequently, the effect of DREG-56 on AGP-induced [Ca2+]i increase was not due to binding of antibodies to Fc receptors. A dose-response study (Fig. 1C
) showed that, in neutrophils pretreated with DREG-56 for 3 min, a detectable Ca2+ response was elicited by exposure to AGP at a dose as low as 0.001 mg/ml [EC50 value 0.009 mg/ml (0.006–0.014)]. By comparison, a shorter incubation with DREG-56 (only 5 s) resulted in a less prominent AGP-induced rise in [Ca2+]i, although it was much more pronounced than the increase provoked by AGP alone. More precisely, the peak rises in [Ca2+]i mediated by AGP were 79 ± 18 and 124 ± 44 nM in cells preincubated with DREG-56 for 5 s and 3 min, respectively; the data are from three different preparations. Longer exposure to DREG-56 (10 min) did not further enhance the AGP-induced Ca2+ response. Another anti-L-selectin antibody (FMC 46; 1 µg/ml) required cross-linking to render the neutrophils highly sensitive to AGP (data not shown). Pretreatment with FMC 46 (1 µg/ml) without cross-linking before the addition of DREG-56 inhibited the subsequent AGP-response (Fig. 2
). Sulfatide is described as a ligand to L-selectin (13
, 15)
. Preincubating neutrophils with sulfatide (10 µg/ml) did not make the cells more sensitive to AGP but, like FMC 46, antagonized DREG-56 enhancement of the AGP response (Fig. 2)
.
|
Other cell adhesion molecules (CAMs) besides L-selectin may mediate intracellular signaling in neutrophils. However, in our control experiments the AGP-evoked elevation of [Ca2+]i in neutrophils was not enhanced by monoclonal antibodies against the β2-integrin subunit CD18 (1 µg/ml) alone (Fig. 2)
or cross-linked (data not shown). Furthermore, replacement of AGP with other human plasma proteins like albumin, transferrin, or fibrinogen did not produce Ca2+ responses in neutrophils; the influence of these proteins was investigated in both the absence and presence of DREG-56 (results for transferrin illustrated in Fig. 2
; data on albumin and fibrinogen not shown). To elucidate the importance of a minor preactivation, we pretreated neutrophils with ATP (10 µM) instead of DREG-56, and the results showed that such prestimulation caused a weak Ca2+ response but did not affect the subsequent AGP-induced increase in [Ca2+]i (Fig. 2)
. On the other hand, strong preactivation with secretory concentrations (10–100 nM) of fMLP resulted in AGP-sensitive neutrophils (Fig. 2)
. This effect was not registered in neutrophils pretreated with a lower chemotactic concentration (1 nM) of fMLP (not shown). Cross-linking of antibodies that bind CAMs often results in a prominent increase in [Ca2+]i in neutrophils (21)
. Theoretically, AGP may act as a cross-linker for DREG-56, and thereby would initiate L-selectin-mediated Ca2+ responses. To examine whether that was the case, we treated neutrophils with DREG-56, then exposed them to a secondary anti-mouse antibody and to AGP. Figure 2
shows that addition of the secondary antibody resulted in a distinct rise in [Ca2+]i in the cells, but did not influence the subsequent AGP-mediated Ca2+ response. Consequently, the observed effect could not be explained by a simple cross-linking between AGP and DREG-56.
HL-60 cells are often used to investigate neutrophil responses and signaling. It is well known that undifferentiated HL-60 cells have no surface expression of L-selectin but that DMSO-differentiated (neutrophil-like) HL-60 cells do express L-selectin (albeit to a lesser extent than seen on neutrophils) (19)
. In our study, AGP had no impact on the [Ca2+]i in DREG-56-treated undifferentiated HL-60 cells, but did cause a rise in [Ca2+]i in differentiated HL-60 cells (Fig. 3
). As with neutrophils, it was necessary to pretreat differentiated HL-60 cells with DREG-56 to achieve an AGP-induced Ca2+ response. The characteristics of undifferentiated HL-60 cells were tested by exposing them to 100 nM fMLP (no Ca2+ response), followed by ATP (a distinct [Ca2+]i increase), whereas differentiated cells responded to fMLP (Fig. 3)
.
|
The AGP-induced increase in [Ca2+]i depends on receptor activation and involves PLC
Receptor-mediated Ca2+ signaling in neutrophils and many other cell types is usually due to activation of phospholipase-C (PLC). We found that the PLC inhibitor U 73122 (2 µM) abolished the rises in [Ca2+]i mediated by AGP and 100 nM fMLP in neutrophils (Fig. 4
).
|
Activation of protein kinase-C (PKC) by phorbol esters such as PMA is associated with receptor desensitization in neutrophils (23)
. We found that treatment of neutrophils with PMA (100 nM) markedly reduced the AGP-evoked increases in [Ca2+]i (Fig. 4)
; pre-exposure to PMA also inhibited the Ca2+-mobilizing action of fMLP (100 nM).
Identification of Siglec-5 as a receptor for AGP
Siglecs constitute a newly characterized group of cell surface receptors, and it is thought that neutrophils express the subtypes Siglec-5 and Siglec-9 (11
, 12
, 24)
. AGP is highly sialylated, and therefore we considered the possibility that Siglecs may serve as receptors for AGP. To test that hypothesis, neutrophil lysate was passed through affinity columns that contained immobilized AGP. Bound proteins were eluted (described under "Affinity chromatography"), then analyzed by Western blot with antibodies against Siglec-5 and Siglec-9 (Fig. 5
). The anti-Siglec-5 antibody resulted in strong staining of a band with a molecular mass of 60–80 kDa in the bound fraction (Fig. 5A
, left panel, lane 2), which is consistent with the previously reported molecular weight of Siglec-5. The Siglec-9 antibody stained a 50 kDa protein in the neutrophil lysate (Fig. 5B
, lane 1), but that band was not recovered by affinity chromatography of the bound fraction, which indicates that Siglec-9 does not interact with AGP. In addition, chromatography on columns containing immobilized desialylated AGP (desialAGP) did not result in binding of Siglec-5 (Fig. 5A
, right panel, lane 2). Furthermore, flow cytometry also showed that isolated neutrophils express Siglec-5 (Fig. 6
). Treatment with DREG-56 (1 µg/ml; 3 min) did not cause changes in surface expression of Siglec-5.
|
|
To study whether Siglec-5 was involved in the Ca2+-mobilizing action of AGP, anti-Siglec-5 antibody was added to neutrophils prior to adding AGP. The addition of anti-Siglec-5 provoked an immediate Ca2+ response in neutrophils that were pre-exposed to DREG-56 (Fig. 8A
). The antibody did not produce any Ca2+ response in the neutrophils that were not pretreated with DREG-56. Furthermore, the anti-Siglec-5 antibody antagonized the AGP-induced rise in [Ca2+]i (Fig. 7
A). Similar effects were obtained when differentiated HL-60 cells were used (Figs. 7B
, 8C)
. AGP N-glycans can present terminal sialic acid in both 2–3 and 2–6 linkage to galactose, and recent reports have indicated that 3'-sialyllactose and 6'-sialyllactose polyacrylamide conjugates can bind to Siglec-5 (24
, 25)
. Consequently, we replaced the anti-Siglec-5 antibody with 100 µg/ml 3' sialyllactose (3'SL) or 6' sialyllactose (6'SL), and the results show that both 3'SL (Fig. 8A
) and 6'SL (not shown) antagonized the AGP-induced [Ca2+]i rise in neutrophils. Similar to the anti-Siglec-5 antibody, the sialyl lactose oligosaccharides also directly provoked a Ca2+ response in neutrophils treated with DREG-56 whereas lactose (100 µg/ml) did not influence the Ca2+ response (data not shown). The importance of sialic acid residues on AGP was further evaluated by treating AGP with neuraminidase or mild periodate. Analysis of the enzyme-treated AGP showed a reduction of 70% in the sialic acid content of neuraminidase-treated AGP (desialAGP). Figure 8
shows that the addition of desial AGP (Fig. 8A
) or mild periodate-treated AGP (Fig. 8B
) produced a significantly smaller rise in [Ca2+]i than that caused by unaltered AGP.
|
|
Src family tyrosine kinases and phosphoinositide 3-kinases (PI3K) are essential for the L-selectin-mediated amplification of the rise in [Ca2+]i induced by AGP/Siglec-5 in neutrophils
Our data suggested that intracellular signaling mediated by L-selectin may be pivotal for AGP/Siglec-5-mediated Ca2+ responses. It was previously proposed that anti-L-selectin (without cross-linking antibodies) activates Src family tyrosine kinases and mitogen-activated protein kinase in neutrophils, and that effect leads to increased protein tyrosine phosphorylation (15
, 26)
.
We evaluated the impact of Src activation on AGP-induced Ca2+ responses by incubating neutrophils with two chemically unrelated Src inhibitors (PP2, 10 or 20 µM and SU 6656, 2 or 20 µM) before exposing the cells to DREG-56, followed by AGP. The results showed that PP2, but not its inert analog PP3, significantly reduced the Ca2+ response mediated by AGP (Fig. 9
A); similar findings were obtained when using the Src inhibitor SU 6656 (data not shown). On the other hand, the p38 mitogen-activated protein kinase (p38 MAPK) inhibitor SB 203580 (10 µM) did not affect the AGP-provoked Ca2+ response (data not shown). We also tested inhibitors against PI3K (wortmannin; 100 nM and LY 294002; 10 µM) and protein tyrosine kinase SYK (Piceatannol; 5 µM) in order to clarify the signaling pathways in DREG-56-enhanced AGP signaling. As seen in Fig. 9B
, inhibition of PI3K abolishes the DREG-56-induced enhancement of AGP Ca2+ signaling, whereas SYK does not seem to be involved.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 3
, 27
, 28)
, but the signaling mechanisms underlying that involvement are not yet known. In this investigation, we found that AGP dose-dependently induced rapid rises in [Ca2+]i in DREG-56-pretreated neutrophils. Notably, the concentration of AGP required for a detectable effect on neutrophils was 500-fold lower than the concentration normally found in plasma. Hence, we suggest that this may represent an important Ca2+-mobilizing action by which AGP modulates neutrophil functional responses and thereby inflammation. Furthermore, addition of human plasma proteins other than AGP (e.g., albumin, transferrin, fibrinogen) to DREG-56-pretreated cells did not cause Ca2+ mobilization. Hence, we conclude that the observed impact of AGP on isolated neutrophils does not represent an unspecific cell-protein interaction. Intriguingly, pretreatment of neutrophils with an antibody against L-selectin (i.e., DREG-56 or cross-linking of FMC 46) was necessary to obtain a marked AGP-induced mobilization of Ca2+. Antibodies against CD18 did not have the same effect. The observed AGP-induced Ca2+ response was not due to cross-linking between AGP and the DREG-56 antibody. More specifically, cross-linking of DREG-56 with a secondary antibody did not change the magnitude of the subsequent Ca2+ response mediated by AGP. The significance of L-selectin was further established by utilizing undifferentiated (L-selectin-negative) and differentiated (L-selectin-positive) HL-60 cells, where DREG-56 combined with AGP caused rises in [Ca2+]i only in the latter cells. L-selectin ligands such as FMC 46 and sulfatide both antagonized DREG-56 enhancement of the AGP-induced Ca2+ response. The present study also reveled that a classical priming stimulus like ATP did not facilitate AGP-induced Ca2+ mobilization. However, a high concentration of fMLP changed the neutrophils to a more AGP-sensitive state. Consequently, the observed effect of AGP is not absolutely restricted to pre-engagement of L-selectin. Taken together, our observations indicate that L-selectin is not the putative AGP receptor, but that pre-engagement of L-selectin is important for unmasking the signaling action of AGP. Strong activation of neutrophils with fMLP can also cause the AGP-sensitive state.
The AGP-induced Ca2+ response exhibited the kinetics typical of receptor-operated signal transduction: a rapid but transient increase in [Ca2+]i. We also found that the Ca2+-elevating effect of AGP was abolished in neutrophils pretreated with a PLC inhibitor or a PKC activator. These characteristics are generally related to receptor-mediated rises in [Ca2+]i that occur in neutrophils (23)
, and thus we focused our efforts on identifying the binding site for AGP and its connection with the mobilization of Ca2+. However, we could not exclude the fact that the antagonizing effect of the PKC activator PMA depends partly on accelerated shedding of L-selectin. Because the glycans on AGP are rich in terminal sialic acid residues, we conducted experiments to address the hypothesis that AGP interacts with Siglec molecules on the surface of the neutrophils. Initially, we observed that Siglec-5, but not Siglec-9, bound to AGP and that Siglec-5 did not interact with desialylated AGP. Second, modified forms of AGP (i.e., AGP treated with neuramidase or mild periodate) produced a significantly smaller rise in [Ca2+]i in neutrophils. Finally, the AGP-induced increase in [Ca2+]i in neutrophils was antagonized by pretreatment of the cells with either an anti-Siglec-5 antibody or sialyllactose oligosaccharides that possess Siglec-5 binding properties (24
, 25)
. Based on these observations, we suggest that AGP acts via sialic acid residues to associate with Siglecs on the surface of neutrophils and that this interaction is coupled to PLC-dependent rises in [Ca2+]i. We also noted that the anti-Siglec-5 antibody, 3'SL, and 6'SL induced immediate Ca2+ responses in neutrophils that had been pretreated with the DREG-56 antibody, which confirms the Ca2+ signaling capacity of Siglec-5. Many glycoproteins, such as transferrin, express terminal sialic acid residues. However, exposing DREG-56-treated neutrophils to transferrin did not cause Ca2+ mobilization, perhaps because transferrin displays fewer sialic acids than AGP does or because the 3-dimensional organization of the glycans on AGP is necessary for interaction with Siglecs and subsequent Ca2+ signaling. Although we cannot exclude the possibility there may be other highly sialylated glycoproteins that can cause a Siglec-dependent increase in [Ca2+]i, at least this is not a general phenomenon associated with all plasma glycoproteins. Together, our results provide strong evidence that an interaction between AGP and Siglec-5 leads to rapid elevation of [Ca2+]i in neutrophils, which implies that the acute-phase protein AGP might be an endogenous ligand for Siglecs on these cells.
Our findings also suggest that engagement of L-selectin facilitates the ability of Siglec-5 to achieve rapid "outside-in" signaling. An alternative explanation could have been that DREG-56 rapidly increases surface expression of Siglec-5, which could make the neutrophils more sensitive to AGP. However, flow cytometry showed that DREG-56 treatment did not change Siglec-5 surface expression (Fig. 5)
. Hence, the DREG-56-induced amplification of the AGP response could not be explained by Siglec-5 up-regulation. It has been shown that L-selectin-induced signaling involves rises in [Ca2+]i and tyrosine phosphorylation, and therefore activation of MAPK and Src tyrosine kinases also (15
, 26
, 29
, 30)
. Based on the results obtained from using ATP, followed by AGP, we conclude that a primary increase in [Ca2+]i does not promote the subsequent AGP-induced rise in [Ca2+]i. Furthermore, drug characterization of the AGP-elicited [Ca2+]i increase did not indicate a role for p38 MAPK. However, we found that the Src inhibitors SU6656 and PP2 (but not the inactive analog PP3), as well as PI3K inhibitors (wortmannin and LY 294002), antagonized the AGP-induced elevation of [Ca2+]i in DREG-56-treated neutrophils. This finding suggests that preactivation of Src tyrosine kinases and PI3K is essential for the AGP-augmented mobilization of Ca2+. A high concentration of fMLP also enhanced the Ca2+ response induced by AGP, and chemotactic peptide-mediated signal transduction has indeed been associated with PI3K and Src activation (31
32
33
34
35
36
37)
.
Activation of Src tyrosine kinases plays a pivotal role in neutrophil intracellular signaling pathways. For instance, Src presumably phosphorylates various substrate proteins, including immunoreceptor tyrosine-based activation motifs (ITAMs) and ITIM domains (38
, 39)
. The ITIMs recruit protein tyrosine phosphatases, which may elicit signaling events that ultimately inhibit leukocyte functions (40)
. The cytoplasmic region of Siglec-5 and other CD33-related Siglecs are thought to comprise ITIM-like domains (11)
. It has been reported that activation of Siglecs impedes agonist-induced Ca2+ influx in rat basophilic leukemia cells (41)
; however, that effect was not found to be associated with phosphorylation of ITIMs. The molecular and cellular actions associated with Siglec molecules are far from understood. For instance, Siglec-5 has been found to enhance the chemotactic peptide-induced respiratory burst (42)
. Furthermore, results that demonstrate Siglec-mediated inhibition of Ca2+ responses and leukocytes functions have been obtained by cross-linking of Siglec antibodies (41)
, an approach that differs vastly from that used in our investigation. In contrast, our study presents strong evidence that interaction between AGP and Siglecs leads to increased [Ca2+]i and that L-selectin plays an important role in that event. We also show that AGP triggers Ca2+ mobilization without measurable delay, which indicates that phosphatase-recruiting ITIMs are not involved. In addition, it has been shown that the intracellular domain of Siglec-5 consists of an SAP binding motif that might serve as another potential phosphorylation site (11)
. SAP binding motifs may also explain the dual effect of Siglec molecules on cytosolic Ca2+ homeostasis in neutrophils, although that assumption must be substantiated. Recently, Angata et al. (43)
identified a new Siglec subtype (Siglec-14) that is almost identical to Siglec-5 with regard to sequence and glycan binding properties. Those investigators noted that Siglec-14 is associated with the activating adaptor protein DAB 12 and therefore can act as an activating receptor. Hence, they proposed that Siglec-5 and Siglec-14 are "paired receptors" with opposite intracellular signaling functions. Consequently, it is possible that the Ca2+ response induced by AGP is associated with Siglec-14 but not with Siglec-5. Nonetheless, to confirm that suggestion, further research is needed to fully establish the concept of paired receptors and to determine whether Siglec-14 is expressed on neutrophils. We suggest that activation of Src tyrosine kinases and PI3K is a key event in the cross-talk between L-selectin and AGP/Siglec-5 signaling, but work is needed to elucidate the precise molecular interactions that underlie the Ca2+ mobilization induced by AGP.
From a physiological perspective, our results indicate that neutrophils circulating in the bloodstream are essentially insensitive to AGP. It is well known that neutrophil rolling on endothelial cells depends partly on L-selectin (44
, 45)
; thus, it is possible that AGP acts as a secondary activator of neutrophils by inducing rises in [Ca2+]i early during the interaction between these cells and the endothelium. Changes in AGP glycosylation during inflammation might constitute a potential regulatory mechanism. It was recently shown that AGP is produced and stored inside neutrophils (46)
, which suggests that AGP may function as an autocrine signaling molecule in these cells and may also participate in modulation of neutrophils in the extravascular space. However, the exact role of AGP in modulating neutrophil responses in vivo remains to be clarified.
In conclusion, we found that AGP has an intrinsic capacity to induce cytosolic Ca2+ responses in neutrophils. Moreover, we suggest that Siglecs (presumably Siglec-5) are surface receptors for AGP, indicating that sialic acid residues are necessary for binding of AGP and the Ca2+ mobilization it induces. Therefore, AGP may be a signaling molecule that is directly involved in regulating neutrophil functions. Notably, we also observed that the signaling capacity of AGP was markedly enhanced by engagement of L-selectin. Furthermore, the AGP-induced Ca2+ response involved PLC, Src, and PI3K activation. Taken together, this suggests an important and intriguing interplay between the two lectins in the regulation of neutrophil responses and may suggest a role for AGP in regulation of the innate immune system.
|
ACKNOWLEDGMENTS |
|---|
Received for publication April 5, 2007. Accepted for publication June 7, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Varki Natural ligands for CD33-related Siglecs? Glycobiology, August 1, 2009; 19(8): 810 - 812. [Full Text] [PDF]
|
|
|
|
 |
|
 A. F. Carlin, S. Uchiyama, Y.-C. Chang, A. L. Lewis, V. Nizet, and A. Varki Molecular mimicry of host sialylated glycans allows a bacterial pathogen to engage neutrophil Siglec-9 and dampen the innate immune response Blood, April 2, 2009; 113(14): 3333 - 3336. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
