Mr. van Rossum and Dr. Fazzini contributed equally to this work.
The prototypic tissue pentraxin PTX3, in contrast to the short pentraxin serum amyloid P, inhibits phagocytosis of late apoptotic neutrophils by macrophages
Article first published online: 5 AUG 2004
Copyright © 2004 by the American College of Rheumatology
Arthritis & Rheumatism
Volume 50, Issue 8, pages 2667–2674, August 2004
How to Cite
van Rossum, A. P., Fazzini, F., Limburg, P. C., Manfredi, A. A., Rovere-Querini, P., Mantovani, A. and Kallenberg, C. G. M. (2004), The prototypic tissue pentraxin PTX3, in contrast to the short pentraxin serum amyloid P, inhibits phagocytosis of late apoptotic neutrophils by macrophages. Arthritis & Rheumatism, 50: 2667–2674. doi: 10.1002/art.20370
- Issue published online: 5 AUG 2004
- Article first published online: 5 AUG 2004
- Manuscript Accepted: 29 MAR 2004
- Manuscript Received: 30 JUN 2003
- Ministero della Salute
- European Union
- Associazione Italiana per la Ricerca sul Cancro
Phagocytosis of apoptotic cells can be facilitated by complement components and short pentraxins, such as serum amyloid P (SAP). In contrast, the long pentraxin PTX3 was shown to inhibit phagocytosis of apoptotic Jurkat cells by dendritic cells and to bind late apoptotic polymorphonuclear leukocytes (PMNs). Recently, levels of the pentraxin PTX3 were shown to parallel disease activity in small-vessel vasculitis, which is often characterized by leukocytoclasia, a persistence of leukocyte remnants in the vessel wall. We undertook this study to test our hypothesis that PTX3 inhibits phagocytosis of late apoptotic PMNs by macrophages, thereby leading to their accumulation in the vessel wall.
Macrophages were allowed to phagocytose late apoptotic or secondary necrotic PMNs that were incubated with or without PTX3 for 30 minutes prior to phagocytosis. Phagocytosis was allowed to occur in the presence of 30% normal human serum with or without SAP and with or without depletion of complement. To discriminate between an inhibitory effect of PTX3 on binding and the internalization of apoptotic PMNs into macrophages, internalization was blocked by cytochalasin B.
SAP and complement were both necessary for effective in vitro phagocytosis. In contrast, PTX3 inhibited phagocytosis in a dose-dependent manner, from 11% inhibition at 6.25 μg/ml to almost complete inhibition at 100 μg/ml. Furthermore, PTX3 partly affected binding of apoptotic PMNs to macrophages.
PTX3, in contrast to SAP and complement, inhibits phagocytosis of late apoptotic PMNs by monocyte-derived macrophages in a dose-dependent manner. Therefore, PTX3 can play a role in the development of leukocytoclasia by affecting the clearance of apoptotic PMNs, thereby inducing their accumulation in the vessel wall.
Pentraxins can be divided into two structural classes: the classic short pentraxins, such as C-reactive protein (CRP) and serum amyloid P (SAP), with monomeric molecular weight of ∼25 kd, and the recently described long pentraxins, with molecular weights of 40–50 kd (for review, see ref. 1). Recently, the long pentraxin PTX3 was identified as a novel acute-phase reactant in active vasculitis. Fazzini et al showed that PTX3 is produced at sites of inflammation, and levels of PTX3 can be used as an independent laboratory indicator for disease activity in small-vessel vasculitis (2). Serum levels of PTX3 only partially corresponded to levels of CRP. Furthermore, in the same study, these investigators showed that activated endothelial cells produce PTX3 at sites of active vasculitis, in contrast to hepatically produced CRP.
Serum concentrations of CRP can increase up to 1,000-fold within a few hours during the acute-phase response (3). In mice, another short pentraxin, SAP, also responds as an acute-phase reactant (3). The function of peripherally produced PTX3 is currently unclear. Recent studies with the pentraxins SAP and CRP have demonstrated an important facilitating role of these molecules in phagocytosis of apoptotic cells (4). PTX3 was shown to be involved in phagocytosis as well. Rovere et al reported that PTX3 specifically binds to apoptotic Jurkat cells and subsequently inhibits their phagocytosis by dendritic cells (DCs) (5). Additionally, they showed that late apoptotic polymorphonuclear leukocytes (PMNs) were able to bind PTX3. In view of the persistence of late apoptotic or secondary necrotic PMNs in active leukocytoclastic vasculitis (6), and given the association between vasculitis disease activity and serum PTX3 levels (2), we hypothesized that PTX3, produced locally by activated endothelial cells, might play a role in the persistence of late apoptotic or secondary necrotic PMNs in vasculitic lesions.
The histopathologic features of small-vessel vasculitis are often designated leukocytoclastic vasculitis (6). Leukocytoclasia (i.e., the accumulation of unscavenged dead neutrophils within the vessel wall) is normally an uncommon phenomenon. Leukocytoclastic lesions are primarily found in the skin, but other organs may be involved as well (7). In small-vessel vasculitis, levels of proinflammatory cytokines, such as tumor necrosis factor α (TNFα) and interleukin-1β (IL-1β), are elevated (8). TNFα and IL-1β can activate endothelial cells, resulting in adhesion and migration of neutrophils. Furthermore, it has been shown that proinflammatory signals induce the production of PTX3 in endothelial cells, and PTX3 was shown to be expressed in endothelial cells and infiltrating leukocytes at sites of active vasculitis (2, 9).
In the present study, we investigated whether PTX3, in contrast to SAP and complement, inhibits the phagocytosis of late apoptotic PMNs by macrophages. Such a finding would explain the phenomenon of leukocytoclasia in small-vessel vasculitis.
MATERIALS AND METHODS
All chemicals used were from Sigma (St. Louis, MO) unless indicated otherwise. Anticoagulant tubes were from BD Vacutainer Systems (Plymouth, UK). Hanks' balanced salt solution (HBSS) and gentamycin were purchased from Gibco Life Technologies (Paisley, UK). RPMI 1640 and fetal calf serum (FCS) were from BioWhittaker Europe (Verviers, Belgium), and culture plates were from Costar (Badhoevedorp, The Netherlands). Human PTX3 was purified from Chinese hamster ovary cells stably and constitutively expressing the protein, as described by Bottazzi et al (10).
Isolation and culture of human neutrophils.
Neutrophils were isolated from blood of healthy human volunteers by Polymorphoprep (Nycomed, Oslo, Norway) density-gradient centrifugation. EDTA-anticoagulated blood (5 ml) was carefully layered over 5 ml of solution and centrifuged at 500g for 30 minutes, and the middle band was harvested. Red blood cells were lysed twice by mixing the cells with 6 ml NH4Cl solution, incubation for 10 minutes on ice, and centrifugation (at 1,200g for 3 minutes). Subsequently, neutrophils were washed in HBSS without calcium and magnesium and centrifuged at 1,200g for 3 minutes. Finally, neutrophils (1 × 106/ml) were reconstituted in RPMI 1640 supplemented with gentamycin (60 μg/ml) and 5% FCS in 6-well plates and aged for 72 hours at 37°C in 5% CO2 to yield late apoptotic or secondary necrotic PMNs. Apoptosis and necrosis were measured by annexin V and propidium iodide staining. For staining, 99 μl binding buffer (10 mM HEPES [pH 7.4], 140 mM NaCl, 5 mM CaCl2), 10 μl propidium iodide (10 μg/ml; Molecular Probes, Leiden, The Netherlands), and 1 μl of fluorescein isothiocyanate–labeled annexin V (Nexins Research, Hoeven, The Netherlands) diluted 1:10 were added to 1 × 106 PMNs of a 100-μl cell suspension. Immediately after incubation for 10 minutes on ice in the dark, immunofluorescence analysis was performed on an Epics-Elite fluorescence-activated cell sorter equipped with a gated amplifier (Coulter Electronics, Mijdrecht, The Netherlands).
Peripheral blood mononuclear cells were isolated by Lymphoprep (Nycomed) density-gradient centrifugation from citrated blood. Healthy controls served as donors. Cells were suspended in medium containing RPMI 1640, gentamycin, and 2% pooled normal human serum (NHS) at a concentration of 1 × 106/ml. Plastic coverslips (13-mm diameter; Nunc, Roskilde, Denmark) were placed in a 16-mm–diameter 24-well plate (Costar, Schiphol, The Netherlands). Cell suspension (0.5 ml) was seeded in every well, and monocytes were subsequently allowed to differentiate into macrophages during 7 days at 37°C in a 5% CO2 atmosphere. On days 2 and 5, the medium was supplemented with 0.5 ml fresh medium.
Phagocytosis assay and scoring.
Coverslips with adherent monocyte-derived macrophages were washed with RPMI 1640 containing 1% NHS to remove nonadherent cells and were transferred into 24-well plates. Late apoptotic PMNs were incubated for 30 minutes with various concentrations of PTX3 (0, 3.125, 6.25, 12.5, 25, 50, and 100 μg/ml) and were resuspended in RPMI 1640 containing 30% NHS. Next, the PMN suspensions (0.3 ml/well) containing 5 × 105 cells were added to the 24-well plates containing monocyte-derived macrophages (5 × 104/well), and cell interaction was allowed for 30 minutes at 37°C in 5% CO2. Cell interaction was also allowed to occur in SAP-depleted and complement-inactivated serum. SAP-depleted serum was made by passing NHS through an agarose column enriched with high-electroendosmosis agarose (lot no. AG 0493; FMC Bioproducts, Rockland, ME), as described previously (4). Complement-inactivated serum was made by heating NHS at 56°C for 30 minutes. Subsequently, coverslips were washed with 0.4% human serum albumin to remove nonphagocytosed cells.
Scoring of phagocytosis was done as described (4, 11). Briefly, macrophages were flattened by centrifugation at 25g for 10 minutes and air-dried. Cells were fixed in ethanol and stained for myeloperoxidase (MPO) as a marker for ingested PMNs. To this end, samples were incubated for 30 minutes with an anti-MPO monoclonal antibody (266.6K2; IQProducts, Groningen, The Netherlands), washed 3 times with phosphate buffered saline, and subsequently incubated with horseradish peroxidase–conjugated goat anti-mouse antibody (Dako, Glostrup, Denmark). Dilutions of antibodies were made according to the manufacturers' protocols. Cells were washed and allowed to react with diaminobenzidine and H2O2. Finally, nuclear staining of monocyte-derived macrophages was performed with hematoxylin (Merck, Darmstadt, Germany). Preparations were scored at 40× magnification by regular light microscopy, and phagocytosis was assessed by counting the number of ingested PMNs per individual macrophage per total of 100 macrophages (phagocytosis index). Only PMNs clearly within the perimeter of the macrophage were counted.
Assessment of binding.
Cytochalasin B (Sigma, Zwijndrecht, The Netherlands) was used to block internalization (but not binding) of apoptotic cells and was added to the 24-well plates with coverslip-adherent monocyte-derived macrophages (5 × 104/well) for 30 minutes prior to interaction with late apoptotic neutrophils. Monocyte-derived macrophages were allowed to interact with apoptotic PMNs for 30 minutes at 37°C in 5% CO2 in the presence of cytochalasin B (24 μg/ml). Scoring was done as described above, except that binding was scored by counting the number of bound PMNs per individual macrophage per total of 100 monocyte-derived macrophages.
Results are expressed as the mean ± SEM of the number of independent experiments. Statistical analysis was performed using Student's unpaired t-test and GraphPad Prism, version 3.0 (GraphPad Software, San Diego, CA).
We generated late apoptotic or secondary necrotic PMNs by aging for 72 hours, since PTX3 specifically binds to late apoptotic PMNs (5). Fluorescence-activated cell sorting analysis showed that a mean ± SEM 82.0 ± 5.2% of cells were positive for annexin V and propidium iodide (Figure 1).
Phagocytosis of late apoptotic PMNs by monocyte-derived macrophages in 30% NHS showed a mean ± SEM phagocytosis index of 45.1 ± 1.2% (Figure 2). Phagocytosis was shown to be complement dependent and partially SAP dependent. Inactivation of complement resulted in a decrease in the phagocytosis index from 45.1 ± 1.2% to 3.3 ± 1.2% (P < 0.0001). Depletion of SAP resulted in a >50% decrease in the phagocytosis index to 21.8 ± 2.1% (P < 0.0001) (Figure 2B).
In order to evaluate the effects of PTX3 on phagocytosis, late apoptotic PMNs were incubated for 30 minutes at room temperature with various concentrations of PTX3 in the presence of 30% NHS. Subsequently, we assayed the phagocytosis of PTX3-incubated PMNs. Incubation of late apoptotic PMNs with 100 μg/ml PTX3 resulted in a significant decrease in the phagocytosis index to 3.6 ± 0.8% (Figure 2). The inhibitory effect of PTX3 on phagocytosis of late apoptotic PMNs was dose dependent (Figure 2C).
Next, to investigate whether the inhibitory effect of PTX3 was due to a defect in binding of late apoptotic PMNs, we incubated macrophages with cytochalasin B. Cytochalasin B–treated cells are incapable of ingesting particles, but they can still bind particles to membrane receptors (12). Cytochalasin B (24 μg/ml) almost completely disrupted internalization, whereas binding was still visible (Figure 3). When PMNs incubated with PTX3 (50 μg/ml) were allowed to interact with cytochalasin B–treated monocyte-derived macrophages, binding was significantly decreased from 76.5 ± 1.5% to 37 ± 2.0% (P < 0.004) (Figure 3C).
Apoptotic cells are specifically recognized and rapidly engulfed by phagocytic cells such as macrophages and DCs. The mechanisms of recognition and removal are complex and incompletely understood. Apoptosis results in a variety of surface changes, such as exposure of phosphatidylserine on the outer membrane of apoptotic cells. In addition, carbohydrates such as fucose and N-acetylglucosamine are increasingly expressed during apoptosis (13, 14). Subsequently, collectins and collectin-like molecules can bind to these newly expressed molecules (15–18). The collectin-like component of complement C1q has been shown to be involved in apoptotic cell recognition (17, 18). C1q can opsonize apoptotic cells and interact with complement receptors such as CR3 and CR4 (19). C1q is thought to play an important role in phagocytosis. Macrophages from C1q-deficient mice have shown a reduced capacity to phagocytose apoptotic thymocytes (19, 20). Binding of complement is a rather late event during apoptotic cell death and is an immediate early feature of necrotic cells. Therefore, complement might serve as an opsonin for late apoptotic or secondary necrotic cells that have escaped normal clearing mechanisms (21). Binding to apoptotic cells has also been demonstrated for the pentraxins (5, 15, 22). The short pentraxins SAP and CRP can bind to apoptotic and necrotic cells, and they can facilitate phagocytosis, possibly by interaction with Fcγ receptors (FcγR) (23).
Recently, Bijl et al demonstrated that SAP binds to late apoptotic Jurkat cells and facilitates their phagocytosis by monocyte-derived macrophages (4). In the present study, we demonstrated that SAP also facilitates phagocytosis of late apoptotic PMNs by monocyte-derived macrophages, since depletion of SAP resulted in a >50% decrease in phagocytosis. Phagocytosis of late apoptotic PMNs was shown to be complement dependent as well. The pentraxin PTX3 also appeared to be involved in phagocytosis (5, 24). Rovere et al (5) demonstrated that PTX3 binds specifically to apoptotic cells and inhibits phagocytosis of apoptotic Jurkat cells by DCs. Binding of PTX3 to apoptotic Jurkat cells was dose dependent and saturable. Furthermore, Rovere et al showed that only late apoptotic neutrophils bind PTX3. For that reason, we generated late apoptotic or secondary necrotic PMNs by aging for 72 hours. PMNs aged for 72 hours stained positive with annexin V and propidium iodide.
PTX3 proved to inhibit phagocytosis of late apoptotic PMNs by monocyte-derived macrophages. This inhibition was dose dependent. To investigate whether PTX3 interfered with the opsonizing effects of SAP, SAP-depleted serum was used. In the presence of SAP-depleted serum, PTX3 still inhibited phagocytosis of late apoptotic PMNs by macrophages. To determine whether this was due to a disturbance in binding of apoptotic cells, we preincubated macrophages with cytochalasin B, which interferes with membrane dynamics and thereby hampers internalization (12). Cytochalasin B–treated macrophages did bind apoptotic PMNs, whereas internalization was almost completely blocked. When PTX3-coated PMNs were added to cytochalasin B–treated macrophages, binding was significantly affected. Binding of PTX3-incubated PMNs was reduced by 50% compared with nonincubated PMNs, which suggests that inhibition of phagocytosis by PTX3 is partly due to reduced binding, whereas the residual inhibitory effect is probably due to effects on internalization.
Rovere et al (5) have suggested that PTX3 did not influence binding of apoptotic cells to immature DCs, but only influenced their internalization. Physicochemical differences in the plasma membranes, the different expression of membrane receptors between macrophages and immature DCs, or the lack of exogenous serum cofactors in the study by Rovere et al may all be relevant to explain these apparent discrepancies. Opsonization of apoptotic cells by classic pentraxins has been suggested to lead to direct or indirect recognition by phagocytic cells. Indirect recognition occurs by activating complement, thereby enabling complement receptor–dependent uptake via CR3 and CR4, whereas direct recognition takes place via FcγR (19). CRP binds to FcγRI and FcγRII, whereas SAP can also interact with FcγRIII (25–27). However, other studies have suggested that CRP does not bind to FcγR (28). Saeland et al (29) indicated that the use of IgG1 anti-CRP monoclonal antibodies to demonstrate binding of CRP to FcγR raises technical problems, since these antibodies can bind through their Fc portion to FcγRII. The use of biotinylated anti-CRP or anti-CRP F(ab′)2 for this purpose did not result in binding of CRP to FcγRII (29). It therefore remains a matter of debate whether CRP really binds to FcγR.
The receptor capable of binding PTX3 has not yet been identified, and the mechanisms by which PTX3 inhibits uptake of apoptotic cells are therefore unclear. In vitro, PTX3 inhibits clearance of apoptotic cells substantially at a concentration of 50 μg/ml, whereas PTX3 in the serum of patients with active untreated vasculitis reaches mean ± SEM levels of only 6.17 ± 4.77 ng/ml (2), which calls into question the relevance of the in vitro findings. However, since PTX3 is locally produced in inflammatory tissue, serum levels of PTX3 do not reflect what is locally present at the site of inflammation.
The biologic significance of the inhibitory capacity of PTX3 is still a subject of speculation. It has been proposed that pentraxin-mediated clearance of apoptotic cells, similar to complement, represents a backup mechanism for the clearance of late apoptotic cells in situations in which apoptotic cell load is high and clearance capacity is low (30). Rovere et al (5) suggested that PTX3 acts as a local regulator inhibiting local inflammatory uptake of late apoptotic cells by immature DCs, thereby preventing antigen presentation by antigen-presenting cells, which can be relevant for preventing induction of autoimmunity.
The inhibitory effect of PTX3 on phagocytosis of apoptotic PMNs may be important in view of leukocytoclasia. Small-vessel vasculitides are histologically characterized by leukocytoclasia (i.e., the accumulation of unscavenged apoptotic or necrotic PMNs in or around the vessel wall). The defective clearance of cell debris can in turn be involved in the maintenance of peripheral inflammation. PTX3, a long extrahepatically produced pentraxin, is released by a variety of cells in vitro, such as fibroblasts, endothelial cells, and cells of the monocytic lineage (9, 31, 32). Since PTX3 can be produced by endothelial cells in active skin lesions of patients with vasculitis (2), we hypothesize that inhibition of the phagocytosis of apoptotic neutrophils by peripherally produced PTX3 is responsible for the phenomenon of leukocytoclasia in small-vessel vasculitis. It therefore seems relevant to demonstrate the presence of PTX3 in lesional tissue from patients with leukocytoclastic vasculitis. Such studies are under way in our laboratory.