Membrane-anchored proteoglycans of mouse macrophages: P388D1 cells express a syndecan-4-like heparan sulfate proteoglycan and a distinct chondroitin sulfate form

Glycosaminoglycans located on neutrophils and monocytes impact on CXCL8- and CCL2-induced cell migration

The role of glycosaminoglycans on the surface of immune cells has so far been less studied compared to their participation in inflammatory responses as members of the endothelium and the extracellular matrix. In this study we have therefore investigated if glycosaminoglycans on immune cells act in concert with GPC receptors (i.e. both being cis-located on leukocytes) in chemokine-induced leukocyte mobilisation. For this purpose, freshly-prepared human neutrophils and monocytes were treated with heparinase III or chondroitinase ABC to digest heparan sulfate -chains or chondroitin sulfate-chains, respectively, from the leukocyte surfaces. Subsequent analysis of CXCL8- and CCL2-induced chemotaxis revealed that leukocyte migration was strongly reduced after eliminating heparan sulfate from the surface of neutrophils and monocytes. In the case of monocytes, an additional dependence of CCL2-induced chemotaxis on chondroitin sulfate was observed. We compared these results with the effect on chemotaxis of a heparan sulfate masking antibody and obtained similarly reduced migration. Following our findings, we postulate that glycosaminoglycans located on target leukocytes act synergistically with GPC receptors on immune cell migration, which is further influenced by glycosaminoglycans located on the inflamed tissue (i.e. trans with respect to the immune cell/GPC receptor). Both glycosaminoglycan localization sites seem to be important during inflammatory processes and could potentially be tackled in chemokine-related diseases.

Syndecan-1 regulates the biological activities of interleukin-34

IL-34 is a challenging cytokine sharing functional similarities with M-CSF through M-CSFR activation. It also plays a singular role that has recently been explained in the brain, through a binding to the receptor protein tyrosine phosphatase RPTPβ/ζ. The aim of this paper was to look for alternative binding of IL-34 on other cell types. Myeloid cells (HL-60, U-937, THP-1) were used as cells intrinsically expressing M-CSFR, and M-CSFR was expressed in TF-1 and HEK293 cells. IL-34 binding was studied by Scatchard and binding inhibition assays, using 125I-radiolabelled cytokines, and surface plasmon resonance. M-CSFR activation was analysed by Western blot after glycosaminoglycans abrasion, syndecan-1 overexpression or repression and addition of a blocking anti-syndecan antibody. M-CSF and IL-34 induced different patterns of M-CSFR phosphorylations, suggesting the existence of alternative binding for IL-34. Binding experiments and chondroitinase treatment confirmed low affinity binding to chondroitin sulphate chains on cells lacking both M-CSFR and RPTPβ/ζ. Amongst the proteoglycans with chondroitin sulphate chains, syndecan-1 was able to modulate the IL-34-induced M-CSFR signalling pathways. Interestingly, IL-34 induced the migration of syndecan-1 expressing cells. Indeed, IL-34 significantly increased the migration of THP-1 and M2a macrophages that was inhibited by addition of a blocking anti-syndecan-1 antibody. This paper provides evidence of alternative binding of IL-34 to chondroitin sulphates and syndecan-1 at the cell surface that modulates M-CSFR activation. In addition, IL-34-induced myeloid cell migration is a syndecan-1 dependent mechanism.

Expression of syndecan-2 in the amoeboid microglial cells and its involvement in inflammation in the hypoxic developing brain

The present study examined the expression of heparan sulphate proteoglycan, syndecan-2 (Sdc-2) in the corpus callosum and the amoeboid microglial cells (AMC) in the neonatal rat brain in response to hypoxia. In 1-day old Wistar rats subjected to hypoxia the mRNA and protein expression of Sdc-2 in the corpus callosum, heavily populated by AMC, was increased up to 3 days after the hypoxic exposure. Immunoexpression of Sdc-2 was localized in AMC as confirmed by double labeling using microglial marker. Primary cultures of microglial cells subjected to hypoxia showed a significant increase in Sdc-2 expression. Application of Sdc-2 to microglial cultures under hypoxia increased the release of tumor necrosis factor-alpha, interleukin-1beta, chemokine (C-C motif) ligand 2 (CCL2), and chemokine (C-X-C motif) ligand 12 (CXCL12) by the microglial cells. Additionally, Sdc-2 enhanced the production of reactive oxygen species (ROS) by microglia subjected to hypoxia. Edaravone [3-methyl-1phenyl-2-pyrazolin-5-one], an antioxidant drug, suppressed the hypoxia- and Sdc-2-induced increased production of cytokines, chemokines, and ROS. In the light of these findings, we suggest that Sdc-2 plays an important role in microglial production of inflammatory cytokines, chemokines, and ROS in hypoxic conditions. In this connection, edaravone suppressed the hypoxia- and Sdc-2-induced increased cytokine and ROS production suggesting its therapeutic potential in ameliorating neuroinflammation.

Decoupled syndecan 1 mRNA and protein expression is differentially regulated by angiotensin II in macrophages

It has been established that syndecan-1 is an important modulator of events relevant to acute tissue repair and chronic injury responses. The current studies were designed to examine syndecan-1 expression during atherosclerotic lesion formation and whether angiotensin II influences syndecan-1 expression in macrophages. ApoE knockout mice maintained on an atherogenic diet were treated for 8 weeks with an infusion of angiotensin II to induce atherosclerosis. Immunohistochemistry was employed to characterize the expression of syndecan-1 in atherosclerotic lesions. Quantitative real-time PCR (QRTPCR) was used to define the role of angiotensin II and responsible signaling pathways involved syndecan-1 expression in RAW264.7 murine macrophages. Protein expression and shedding were characterized by fluorescence activated cell sorting (FACS) and slot blot analysis. Syndecan-1 was abundantly expressed in macrophages located within early atherosclerotic lesions. Accordingly, we hypothesized that angiotensin II regulates syndecan-1 expression in macrophages. A time- and dose-dependent study was performed in RAW264.7 macrophages. QRTPCR demonstrated maximum syndecan-1 mRNA up-regulation at 6 h after 500 nM AgII stimulation (threefold; P < 0.05). Through administration of specific inhibitors, we established that ERK/MAPK, PI3K and JNK signaling pathways mediated this effect. FACS and slot blot analyses demonstrated that cAMP induced posttranscriptional syndecan-1 protein expression in a dose-dependent manner with or without initial angiotensin II stimulation. In particular, angiotensin II induced an increase in cell surface syndecan-1 (mean fluorescence intensity: 147 +/- 5.7 vs. 176 +/- 4.8; P < 0.05; n = 3) and accelerated syndecan-1 shedding. Angiotensin II is a potent regulator of syndecan-1 expression in atherosclerotic lesions via a specific effect on macrophages that is mediated by ERK/MAPK, PI3K, and JNK signaling pathways.

Heparan sulfate proteoglycan-dependent neutrophil chemotaxis toward PR-39 cathelicidin

Cathelicidins are mammalian proteins containing a C-terminal cationic antimicrobial domain. Porcine PR-39 cathelicidin affects leukocyte biology. Mechanisms of action may involve alteration of heparan sulfate proteoglycan-dependent functions in inflammatory cells. It was tested whether PR-39 affects human neutrophil migration and if such effects involve heparan sulphate proteoglycans. Neutrophils were from forearm venous blood of healthy donors. Migration was tested in modified Boyden chamber assays. Involvement of heparan sulfate proteoglycans was tested by their chemical modification and by the use of specific antibodies. PR-39 induced migration in neutrophils in a concentration dependent manner. Modification of heparan sulfate proteoglycans with sodium chlorate inhibited migration whereas chemotaxis toward the chemoattractant formyl-Met-Leu-Phe was not affected. Removal of heparan sulfates or chondroitin sulfates from the surface of neutrophils by heparinase or chondroitinase inhibited migration toward PR-39. In conclusion, antimicrobial PR-39 stimulates human neutrophil chemotaxis in a heparan sulfate proteoglycan-dependent manner. Involvment of syndecans is likely as both heparinase and chondroitinase were abrogating. Data suggest active participation of heparan sulfate proteoglycans of neutrophils in cathelicidin peptide-mediated regulation of the antimicrobial host defense.

The role of heparan sulphate in inflammation

The polysaccharide heparan sulphate is ubiquitously expressed as a proteoglycan in extracellular matrices and on cell surfaces. Heparan sulphate has marked sequence diversity that allows it to specifically interact with many proteins. This Review focuses on the multiple roles of heparan sulphate in inflammatory responses and, in particular, on its participation in almost every stage of leukocyte transmigration through the blood-vessel wall. Heparan sulphate is involved in the initial adhesion of leukocytes to the inflamed endothelium, the subsequent chemokine-mediated transmigration through the vessel wall and the establishment of both acute and chronic inflammatory reactions.

Cell surface proteoglycan expression during maturation of human monocytes-derived dendritic cells and macrophages

Cell surface proteoglycans play an important part in the functional and metabolic behaviour of leucocytes. We studied the expression of cell surface proteoglycans in human monocytes, in monocyte-derived immature and mature dendritic cells and in macrophages by metabolic labelling with [(35)S]-sulphate, reverse transcriptase-polymerase chain reaction (RT-PCR) and Western blotting. Immature dendritic cells had the highest metabolic activity for the synthesis of cell surface proteoglycans. The major part of these proteoglycans was in phosphatidylinositol-anchored form and was released after treatment with phospholipase C. A minor part was released by trypsin. Digestion with chondroitinase ABC and mild HNO(2) treatment showed that cell surface proteoglycans had a higher proportion of chondroitin sulphate, both in the phospholipase C and trypsin fractions, suggesting that at least some glypicans contained chondroitin sulphate chains. RT-PCR detected the transcripts of glypicans 1, 3, 4 and 5 and all syndecans. Immature dendritic cells expressed a most complex spectrum of glypicans and syndecans, glypican-1 and syndecan-1 being expressed preferentially by this type of cells. Mature dendritic cells expressed glypican-3, which was not present in other lineages. These results suggest that different mononuclear cells synthesize cell surface proteoglycans actively with characteristic expression of different syndecans and glypicans genes, depending on the degree of cell differentiation and/or maturation.

Heparan sulfate proteoglycans in glomerular inflammation

Heparan sulfate proteoglycans (HSPGs) are glycoproteins consisting of a core protein to which linear heparan sulfate side chains are covalently attached. These heparan sulfate side chains can be modified at different positions by several enzymes, which include N-deacetylases, N- and O-sulfotransferases, and an epimerase. These heparan sulfate modifications give rise to an enormous structural diversity, which corresponds to the variety of biologic functions mediated by heparan sulfate, including its role in inflammation. The HSPGs in the glomerular basement membrane (GBM), perlecan, agrin, and collagen XVIII, play an important role in the charge-selective permeability of the glomerular filter. In addition to these HSPGs, various cell types express HSPGs at their cell surface, which include syndecans, glypicans, CD44, and betaglycan. During inflammation, HSPGs, especially heparan sulfate, in the extracellular matrix (ECM) and at the surface of endothelial cells bind chemokines, which establishes a local concentration gradient recruiting leukocytes. Endothelial and leukocyte cell surface HSPGs also play a role in their direct adhesive interactions via other cell surface adhesion molecules, such as selectins and beta2 integrin. Activated leukocytes and endothelial cells exert heparanase activity, resulting in degradation of heparan sulfate moieties in the ECM, which facilitates leukocyte passage into tissues and the release of heparan sulfate-bound factors. In various renal inflammatory diseases the expression of agrin and GBM-associated heparan sulfate is decreased, while the expression of CD44 is increased. Heparan sulfate or heparin preparations affect inflammatory cell behavior and have promising therapeutic, anti-inflammatory properties by preventing leukocyte adhesion/influx and tissue damage.

Syndecan-4-dependent migration of human eosinophils

BACKGROUND

Heparan sulphate proteoglycans (HSPGs) are important participants in cell surface signalling and critical in controlling cell behaviour. They modulate inflammatory cell maturation and activation, leucocyte rolling, adhesion to endothelium as well as extravasation and chemotaxis. Whether eosinophil's function is affected has not yet been reported.

OBJECTIVE

We investigated the effects of transgenic, recombinant anti-thrombin III and Kybernin P, an anti-thrombin III concentrate, as HSPG ligands on spontaneous and chemokine-triggered migration of normal eosinophils from human peripheral blood in modified Boyden chamber micropore filter assays.

METHODS

Eosinophils from human peripheral blood were purified using magnetic antibody cell sorting. The signalling mechanisms required for anti-thrombin-dependent migration were studied using signalling enzyme blockers. Expression of HSPG core protein mRNA was studied by PCR.

RESULTS

Pre-treatment of eosinophils with anti-thrombin III inhibited chemotaxis toward optimal concentrations of eotaxin or RANTES (regulated upon activation normal T cell expressed and activated). In the absence of the chemokines, direct exposure to gradients of anti-thrombin III stimulated eosinophil migration. The effects of anti-thrombin III were abolished by pre-treating cells with heparinase-1, chondroitinase, sodium chlorate and anti-syndecan-4 antibodies. Syndecan-4 gene expression in eosinophils was confirmed in PCR. In the presence of pentasaccharide, anti-thrombin III lost its effect on the cells. Functional responses were also abrogated by inhibition of protein kinase C, phosphatidylinositol 3-kinase and phosphodiesterase.

CONCLUSION

Data indicate that anti-thrombin III affects eosinophil motility via the effects of its heparin-binding site on cell surface syndecan-4. Ligation of syndecan-4 with anti-thrombin III induces eosinophil migration and deactivates motility toward chemokines. These observations suggest that syndecan-4-dependent signalling may control eosinophil locomotion.

Increased syndecan expression following myocardial infarction indicates a role in cardiac remodeling

Finsen, Alexandra Vanessa, Per Reidar Woldbaek, Jian Li, Jiaping Wu, Torstein Lyberg, Theis Tonnessen, and Geir Christensen. Increased syndecan expression following myocardial infarction indicates a role in cardiac remodeling. Physiol Genomics 16: 301-308, 2004. First published November 18, 2003; 10.1152/physi-olgenomics. 00144.2002.—The purpose of this study was to identify essential genes involved in myocardial growth and remodeling following myocardial infarction (MI). Left ventricular noninfarcted tissues from six mice subjected to MI under general anesthesia and from six sham-operated mice were obtained 1 wk after primary surgery and analyzed by means of cDNA filter arrays. Out of a total of 1,176 genes, 641 were consistently expressed, twenty-three were upregulated and thirteen downregulated. Five genes were only expressed following MI. Syndecan-3, a transmembranous heparan sulfate proteoglycan, was found to be upregulated together with a transcriptional activator of syndecans, Wilms tumor protein 1 (WT-1). Northern blotting demonstrated a significant upregulation of syndecan-1, -2, -3, and -4, WT-1, fibronectin, and basic fibroblast growth factor (FGF) receptor 1. Furthermore, Western blot analysis showed statistically significant increases in protein levels for syndecan-3 and -4. In conclusion, we have identified a subset of genes with increased expression in noninfarcted left ventricular tissue following MI, including syndecans 1–4, WT-1, fibronectin, collagen 6A, and FGF receptor 1. Since the syndecans link the cytoskeleton to the extracellular matrix and function as required coreceptors for FGF, we suggest a role for the syndecans in cardiac remodeling following MI.

Secretion of macrophage urokinase plasminogen activator is dependent on proteoglycans

The importance of proteoglycans for secretion of proteolytic enzymes was studied in the murine macrophage cell line J774. Untreated or 4beta-phorbol 12-myristate 13-acetate (PMA)-stimulated macrophages were treated with hexyl-beta-d-thioxyloside to interfere with the attachment of glycosaminoglycan chains to their respective protein cores. Activation of the J774 macrophages with PMA resulted in increased secretion of trypsin-like serine proteinase activity. This activity was completely inhibited by plasminogen activator inhibitor 1 and by amiloride, identifying the activity as urokinase plasminogen activator (uPA). Treatment of both the unstimulated or PMA-stimulated macrophages with xyloside resulted in decreased uPA activity and Western blotting analysis revealed an almost complete absence of secreted uPA protein after xyloside treatment of either control- or PMA-treated cells. Zymography analyses with gels containing both gelatin and plasminogen confirmed these findings. The xyloside treatment did not reduce the mRNA levels for uPA, indicating that the effect was at the post-translational level. Treatment of the macrophages with xylosides did also reduce the levels of secreted matrix metalloproteinase 9. Taken together, these findings indicate a role for proteoglycans in the secretion of uPA and MMP-9.

Syndecans in inflammation

Cell surface heparan sulfate (HS) influences a multitude of molecules, cell types, and processes relevant to inflammation. HS binds to cell surface and matrix proteins, cytokines, and chemokines. These interactions modulate inflammatory cell maturation and activation, leukocyte rolling, and tight adhesion to endothelium, as well as extravasation and chemotaxis. The syndecan family of transmembrane proteoglycans is the major source of cell surface HS on all cell types. Recent in vitro and in vivo data suggest the involvement of syndecans in the modulation of leukocyte-endothelial interactions and extravasation, the formation of chemokine and kininogen gradients, participation in chemokine and growth factor signaling, as well as repair processes. Thus, the complex role of HS in inflammation is reflected by multiple functions of its physiological carriers, the syndecans. Individual and common functions of the four mammalian syndecan family members can be distinguished. Recently generated transgenic and knockout mouse models will facilitate analysis of the individual processes that each syndecan is involved in.

Inhibition of leukocyte and platelet aggregation in vitro by antithrombin

Glycosaminoglycan-mediated aggregation of cells occurs through adhesion mechanisms in which heparan sulfate chains bind to counter receptors on these cells. As antithrombin interacts with heparan sulfate proteoglycans through its heparin-binding domain and inhibits leukocyte adhesion in ischaemia/reperfusion, it may affect leukocyte aggregation. Leukocyte aggregation was therefore monitored in vitro as the increase in transmission of light through stirred suspensions in a platelet aggregometer. Aggregation curves were quantified as the area under the curve in the first 6 min following stimulation. Leukocytes in platelet-rich plasma were obtained from heparinized whole blood of healthy donors by centrifugation; the ratio of leukocytes to platelets was about 1:45, and the final concentration of autologous plasma was 80%. Neutrophils were purified by dextran sedimentation, density centrifugation, and hypotonic lysis of erythrocytes. Aggregation was induced by phytohaemaglutinin (0.24 mg/mL) or formyl-Met-Leu-Phe (0.2 x 10(-6) M), with or without various concentrations of antithrombin. During the observation period (6 min) no aggregation of leukocytes in platelet-rich plasma or isolated neutrophils could be induced either with medium or with antithrombin (0.2 x 10(0) to 0.2 x 10(-6) IU/mL). Addition of phytohaemagglutinin or formyl-Met-Leu-Phe stimulated aggregation of leukocytes in platelet-rich plasma and neutrophils to different extents. Additional presence of anti-thrombin significantly inhibited phytohaemagglutinin-induced aggregation of leukocytes in platelet-rich plasma, whereas formyl-Met-Leu-Phe-induced aggregation was not affected by antithrombin. Data show that in the presence of plasma and platelets, aggregation of normal white blood cells after stimulation with lectin but not with chemotaxin is inhibited by antithrombin, suggesting specific interactions of antithrombin with lectin-activated processes of neutrophil aggregation that occur in the presence of platelets and/or plasma.

Proteoglycans mediate malaria sporozoite targeting to the liver

Malaria sporozoites are rapidly targeted to the liver where they pass through Kupffer cells and infect hepatocytes, their initial site of replication in the mammalian host. We show that sporozoites, as well as their major surface proteins, the CS protein and TRAP, recognize distinct cell type-specific surface proteoglycans from primary Kupffer cells, hepatocytes and stellate cells, but not from sinusoidal endothelia. Recombinant Plasmodium falciparum CS protein and TRAP bind to heparan sulphate on hepatocytes and both heparan and chondroitin sulphate proteoglycans on stellate cells. On Kupffer cells, CS protein predominantly recognizes chondroitin sulphate, whereas TRAP binding is glycosaminoglycan independent. Plasmodium berghei sporozoites attach to heparan sulphate on hepatocytes and stellate cells, whereas Kupffer cell recognition involves both chondroitin sulphate and heparan sulphate proteoglycans. CS protein also interacts with secreted proteoglycans from stellate cells, the major producers of extracellular matrix in the liver. In situ binding studies using frozen liver sections indicate that the majority of the CS protein binding sites are associated with these matrix proteoglycans. Our data suggest that sporozoites are first arrested in the sinusoid by binding to extracellular matrix proteoglycans and then recognize proteoglycans on the surface of Kupffer cells, which they use to traverse the sinusoidal cell barrier.

Synthesis and characterization of biologically functional biotinylated RANTES

Development of specifically labeled chemokines that retain their biological properties should be useful for analyzing their mechanisms of action both under physiological and pathological conditions. Here, we report the chemical synthesis and characterization of RANTES (regulated upon activation normal T cell expressed and secreted) derivatives that were biotinylated at residues 1, 25, 33, 45, or 67. Gel filtration and ultracentrifugation experiments showed that biotinylation at position 45 or 67 decreased the aggregation tendency of the chemokine to a dimeric state. Competition experiments, using a stably transfected CHO-K1 cell line overexpressing human CCR5, a RANTES receptor, indicated that derivatives biotinylated at positions 1, 25, and 67 bound to CCR5 with the same affinity as native RANTES. Flow cytometry analysis showed that RANTES biotinylated at residue 67 (B67-RANTES) bound more efficiently to primary macrophages than the other derivatives. Such binding was dependent on cell surface glycosaminoglycans (GAGs) since it was reduced when macrophages or HeLa cells expressing or not CCR5 were first treated with GAG-specific enzymes. In addition, B67-RANTES modulated CCR5 expression on lymphocytes and elicited chemotaxis of monocytes in the same manner as unmodified RANTES. Thus, B67-RANTES acts as a CCR5 agonist and may be useful to study the role of RANTES in pathologies such as, for example, human immunodeficiency virus (HIV) infection and inflammatory disorders.

Syndecan-4 mediates antithrombin-induced chemotaxis of human peripheral blood lymphocytes and monocytes

Antithrombin inhibits chemokine-induced migration of neutrophils by activating heparan sulfate proteoglycan-dependent signaling. Whether antithrombin affects migration of other types of leukocytes is not known. We investigated the effects of antithrombin on spontaneous and chemokine-triggered migration of lymphocytes and monocytes from human peripheral blood in modified Boyden chamber micropore filter assays. Lymphocyte and monocyte populations from human peripheral blood were purified using magnetic antibody cell sorting. The signaling mechanisms required for antithrombin-dependent migration were studied using signaling enzyme blockers. Expression of heparan sulfate proteoglycan core protein was studied by RT-PCR and flow cytometry. The antithrombins used were Kybernin®P from human plasma and a monoclonal-antibody-purified preparation from this plasma. Pretreatment of lymphocytes and monocytes with antithrombin inhibited chemotaxis toward optimal concentrations of interleukin-8 or Rantes (regulated upon activation normal T-cell expressed and activated) at concentrations of antithrombin as low as 10 nU/ml. In the absence of the chemokines, direct exposure of cells to gradients of antithrombin stimulated migration. Effects of antithrombin were abolished by pretreating cells with heparinase-1, chondroitinase, sodium chlorate and anti-syndecan-4 antibodies. Expression of syndecan-4 mRNA and protein in monocytes and lymphocytes was demonstrated in RT-PCR and anti-syndecan-4 immunoreactivity assays, respectively. In the presence of pentasaccharide, antithrombin lost its effect on cells. Data indicate that antithrombin directly inhibits chemokine-stimulated migration of monocytes and lymphocytes via the effects of its heparin-binding site on cell surface syndecan-4 by activation of protein kinase C and Rho signaling.

Syndecan-4 as antithrombin receptor of human neutrophils

Antithrombin inhibits chemokine-induced migration of neutrophils by activating heparan sulfate proteoglycan-dependent signaling. Mechanisms of antithrombin's effects on neutrophils were, therefore, studied by testing function and expression of heparan sulfate proteoglycans in RT-PCR or flow cytometry and cell migration assays, respectively. In vitro effects of antithrombin on human neutrophil migration in modified Boyden chambers were abolished by pretreating cells with heparinase-1, chondroitinase, sodium chlorate, and anti-syndecan-4 antibodies. Expression of syndecan-4 mRNA and protein in neutrophils was demonstrated in RT-PCR and anti-syndecan-4 immunoreactivity assay, respectively. In the presence of pentasaccharide, antithrombin lost its activity on the cells. Data suggest that antithrombin regulates neutrophil migration via effects of its heparin-binding site on cell surface syndecan-4.

Inducible expression of the cell surface heparan sulfate proteoglycan syndecan-2 (fibroglycan) on human activated macrophages can regulate fibroblast growth factor action

Monocyte/macrophages play important roles in regulating tissue growth and angiogenesis through the controlled release of heparin-binding growth factors such as fibroblast growth factor (FGF), vascular endothelial growth factor, and heparin binding epidermal growth factor. The action of these potent growth mediators is known to be regulated by adsorption to heparan sulfate proteoglycans (HSPGs) on the surface and within the extracellular matrix of other neighboring cells, which respectively promote or restrict interactions with their signal-transducing receptors on target cells. Here we report on the nature of HSPGs inducibly expressed on the surface of macrophages that confer these cells with the capacity to regulate endogenous growth factor activity. We reveal that activated human macrophages express only a single major 48-kDa cell surface HSPG, syndecan-2 (fibroglycan) as the result of de novo RNA and protein synthesis. In addition, we demonstrate this macrophage HSPG selectively binds the macrophage-derived growth factors FGF-2, vascular endothelial growth factor and heparin binding EGF and can present FGF-2 in a form that transactivates receptor-bearing BaF32 cells. These results define a novel and unique proteoglycan profile for macrophages and imply a key role for syndecan-2 in the delivery of sequestered growth factors by inflammatory macrophages for productive binding to their appropriate target cells in vivo.

Characterization of low-density lipoprotein uptake by murine macrophages exposed to Chlamydia pneumoniae

Exposure to Chlamydia pneumoniae is correlated with atherosclerosis in a variety of clinical and epidemiological studies, but how the organism may initiate and promote the disease is poorly understood. One pathogenic mechanism could involve modulation of macrophage function by C. pneumoniae. We recently demonstrated that C. pneumoniae induces macrophages to accumulate excess cholesterol and develop into foam cells, the hallmark of early atherosclerotic lesions. To determine if C. pneumoniae-induced foam cell formation involved increased uptake of low-density lipoprotein (LDL), the current study examined macrophage association of a fluorescent carbocyanine (DiI)-labeled LDL following infection. C. pneumoniae enhanced the association of DiI-LDL with macrophages in a dose-dependent manner with respect to both C. pneumoniae and DiI-LDL. Interestingly, increased association was inhibited by native LDL and occurred in the absence of oxidation byproducts and in the presence of antioxidants. However, enhanced DiI-LDL association occurred without the participation of the classical Apo B/E native LDL receptor, since C. pneumoniae increased DiI-LDL association and induced foam cell formation in macrophages isolated from LDL-receptor-deficient mice. Surprisingly, DiI-LDL association was inhibited not only by unlabeled native LDL but also by high-density lipoprotein, very low density lipoprotein, and oxidized LDL. These data indicate that exposure of macrophages to C. pneumoniae increases the uptake of LDL and foam cell formation by an LDL-receptor-independent mechanism.

Macrophage released proteoglycans are involved in cell-mediated aggregation of LDL

Aggregated low density lipoprotein (LDL) is taken up by macrophages at enhanced rate, leading to macrophage cholesterol accumulation and foam cell formation. Since macrophages were shown to mediate self aggregation of modified forms of LDL, we sought to study the effect of macrophages on the susceptibility of native LDL to aggregation. Incubation of LDL (100 microg of protein/ml) with J-774A.1 macrophage-like cell line for 18 h at 37 degrees C, led to a 114 and 56% enhanced susceptibility of LDL to aggregation by vortexing and by Bacillus cereus SMase respectively. Macrophage conditioned media (MCMs) that were obtained from J-774A.1 cells also enhanced the susceptibility of LDL to aggregation by vortexing and SMase by 134 and 75% respectively, suggesting the involvement of macrophage secretory products in the enhanced aggregation of LDL. As proteoglycans were shown to be involved in lipoprotein aggregation, we analyzed the possible involvement of macrophage-released proteoglycans in LDL aggregation. Incubation of LDL (100 microg protein/ml) with 25 microg of proteoglycans that were isolated from MCM led to a dose-dependent enhanced susceptibility of LDL to aggregation by vortexing or by SMase by up to 62 and 77% respectively. The stimulatory effect of the MCMs on LDL aggregation was markedly reduced upon MCMs treatment with the glycosaminoglycan hydrolyzing enzyme chondroitinase ABC, chondroitinase AC, but not heparinase. On the contrary, incubation of LDL (100 microg of protein/ml) with increasing concentrations (up to 50 microg/ml) of chondroitin sulfate, or heparan sulfate enhanced the susceptibility of LDL to aggregation by up to 98 or by only 18% respectively, in comparison with non-treated LDL. Since macrophages under atherogenic conditions (cholesterol-loading, cellular lipid peroxidation and activation) demonstrate enhanced secretion of proteoglycans, we finally studied the effect of J-774A.1 macrophages on the susceptibility of native LDL to aggregation under the above atherogenic conditions. Incubation of LDL with cholesterol-loaded macrophages led to a 62% enhanced susceptibility of LDL to undergo aggregation by vortexing, in comparison with LDL that was incubated with non-loaded cells. Macrophage activation with phorbol myristate acetate (5 microM of PMA) also significantly increased cell-mediated aggregation of LDL by 50%, in comparison with non-activated cells. Lipid peroxidized macrophages obtained by cell treatment with either FeSO4 (50 microM), or angiotensin II (10(-7) M) enhanced the susceptibility of LDL to aggregation by 22 or by 39% respectively. These results suggest that under atherogenic conditions, macrophages release proteoglycans, and mainly chondroitin sulfate, which can contribute to cell-mediated formation of aggregated LDL, a potent inducer of macrophage foam cells which are the hallmark of early atherogenesis.

The inhibitory effect of RANTES on the infection of primary macrophages by R5 human immunodeficiency virus type-1 depends on the macrophage activation state

We investigated whether culture conditions could affect the RANTES antiviral effect on human immunodeficiency virus type 1 (HIV-1) infection of primary macrophages. Monocyte-derived macrophages (MDM) were obtained either as (1) the adherent cells of 5-day cultures of blood mononuclear cells (PBMC), followed by 2 days without nonadherent PBMC or added cytokines (MDM-5d), or (2) as the adherent cells recovered from 1-h incubation of PBMC, which were cultured for 7 days with macrophage colony-stimulating factor (M-CSF; MDM-MCSF). Infection of MDM-5d from different donors with HIV-1 R5 strains was reproducibly inhibited by RANTES (IC50 < or = 10 nM), but infection of MDM-MCSF was not inhibited by > or = 100 nM RANTES, even when added at initiation of cultures, although it was still inhibited by a CD4 antibody. RANTES had no antiviral effect when MDM-5d were treated with physiological concentrations of M-CSF or GM-CSF before infection. CCR5 and CXCR4 expression as well as that of other cell surface molecules, including adhesion molecules, was not affected by the cytokines. MDM-MCSF from delta 32CCR5 homozygous individuals did not render them permissive to HIV-1, suggesting that it is unlikely that the virus uses another coreceptor. RANTES binding to MDM was chondroitin sulfate, but not heparan sulfate, dependent, and RANTES bound more efficiently to MDM-5d than to MDM-MCSF. Chondroitin sulfate removal partially offset the RANTES antiviral effect for MDM-5d. Thus RANTES anti-HIV-1 activity for primary macrophages depends on culture conditions and their consequent activation status, which may lead to differences in proteoglycan surface expression. These data may be relevant for the development of chemokine-based therapy for HIV-1 infection.

Serglycin expression during monocytic differentiation of U937-1 cells

Serglycin is the major proteoglycan in most hematopoietic cells, including monocytes and macrophages. The monoblastic cell line U937-1 was used to study the expression of serglycin during proliferation and differentiation. In unstimulated proliferating U937-1 cells serglycin mRNA is nonconstitutively expressed. The level of serglycin mRNA was found to correlate with the synthesis of chondroitin sulfate proteoglycan (CSPG). The U937-1 cells were induced to differentiate into different types of macrophage-like cells by exposing the cells to PMA, RA, or VitD3. These inducers of differentiation affected the expression of serglycin mRNA in three different ways. The initial upregulation seen in the normally proliferating cells was not observed in PMA treated cells. In contrast, RA increased the initial upregulation, giving a reproducible six times increase in serglycin mRNA level from 4 to 24 h of incubation, compared to a four times increase in the control cells. VitD3 had no effect on the expression of serglycin mRNA. The incorporation of (35S)sulfate into CSPG decreased approximately 50% in all three differentiated cell types. Further, the (35S)CSPGs expressed were of larger size in PMA treated cells than controls, but smaller after RA treatment. This was due to the expression of CSPGs, with CS-chains of 25 and 5 kDa in PMA and RA treated cells, respectively, compared to 11 kDa in the controls. VitD3 had no significant effect on the size of CSPG produced. PMA treated cells secreted 75% of the (35S)PGs expressed, but the major portion was retained in cells treated with VitD3 or RA. The differences seen in serglycin mRNA levels, the macromolecular properties of serglycin and in the PG secretion patterns, suggest that serglycin may have different functions in different types of macrophages.

Proteoglycans in macrophages: characterization and possible role in the cellular uptake of lipoproteins

The murine macrophage cell line J774 was incubated with [35S]sulphate. The cell-associated 35S-labelled macromolecules were shown to be proteoglycans and glycosaminoglycans in similar amounts. The possible presence of cell-surface proteoglycans was investigated by incubating [35S]sulphate-labelled cells with trypsin for 15 min. The released material contained approx. 70% free glycosaminoglycan chains and 30% proteoglycans. The latter component was demonstrated by HNO2 treatment to contain heparan sulphate. In the total cell fraction not treated with trypsin a small but significant portion was shown to be chondroitin sulphate proteoglycan. The cell-associated glycosaminoglycans contained both chondroitin sulphate and heparan sulphate. To investigate possible biological functions of cell-surface proteoglycans in macrophages, cells were incubated with NaClO3 to inhibit sulphation of proteoglycans and beta-d-xyloside to abrogate proteoglycan expression. The uptake of oxidized 125I-tyraminylcellobiose-labelled low-density lipoprotein (125I-TC-LDL) was typically two to three times higher than that of native 125I-TC-LDL in untreated J774 cells. The cellular uptake at 37 degreesC of native 125I-TC-LDL was decreased 25% after both NaClO3 and xyloside treatment, whereas the uptake of oxidized 125I-TC-LDL was decreased 35% after both types of treatment. The mRNA levels for the scavenger receptor A-II and the LDL receptor were not affected by NaClO3 or xyloside treatment. Furthermore, fluid-phase endocytosis, measured as uptake of horseradish peroxidase, and receptor-mediated endocytosis, measured as uptake of 125I-TC-ovalbumin, were not affected by NaClO3 treatment of J774 cells. Removal of cell-surface chondroitin sulphate with chondroitinase ABC decreased only the binding of native 125I-TC-LDL, whereas removal of heparan sulphate with heparitinase decreased the binding of both oxidized and native 125I-TC-LDL. Addition of lipoprotein lipase increased the uptake of oxidized 125I-TC-LDL 1.7 times and the uptake of native 125I-TC-LDL 2.1 times. The binding of the former was more sensitive to NaClO3 treatment than the latter. The results presented support the notion that some of the uptake pathways for lipoproteins in the foam-cell-forming macrophages depend on the presence of cell-surface heparan sulphate and chondroitin sulphate.

Role of macrophage glycosaminoglycans in the cellular catabolism of oxidized LDL by macrophages

Macrophage binding sites for oxidized LDL (Ox-LDL) include class A scavenger receptors (SR-As), the CD-36 molecule, and an additional but hitherto unidentified binding site. Because cell-surface glycosaminoglycans (GAGs) were previously shown to be involved in the cellular uptake of native LDL and lipoprotein(a), several strategies to assess the participation of heparan sulfate (HS) and chondroitin sulfate (CS) in macrophage catabolism of Ox-LDL were used. First, incubation of J-774 A.1 macrophage-like cells with either heparinase or chondroitinase, or with both enzymes together, reduced the binding, uptake, and degradation of 125I-Ox-LDL by 20% to 45%, in comparison with control nontreated cells, while catabolism of 125I-labeled acetylated LDL (Ac-LDL) and native LDL were unaffected. Second, the proteoglycan (PG) cellular content was increased by cell enrichment with exogenous GAGs or by using human monocyte-derived macrophages from two patients with Sanfilippo mucopolysaccharidosis, which are characterized by cellular HS accumulation. In these macrophages, cellular uptake of 125I-Ox-LDL increased, while catabolism of 125I-Ac-LDL and native LDL were unaffected. Experiments using conditioned media from control, heparinase-digested, or chondroitinase-digested macrophages indicated that neither secreted GAGs nor released digestion products played any role in Ox-LDL catabolism. To evaluate potential interactions between cell-surface GAGs and known receptors for Ox-LDL, we used excess unlabeled Ac-LDL to block SR-As or anti-CD-36 antibodies to block CD-36, and then examined the catabolism of 125I-Ox-LDL by GAG-enriched or -depleted macrophages. Both excess unlabeled Ac-LDL and anti-CD-36 antibodies reduced 125I-Ox-LDL catabolism, but only excess unlabeled Ac-LDL completely abolished the increase in 125I-Ox-LDL catabolism on GAG enrichment of the cells, indicating a cooperation between exogenous GAGs and cell-surface SR-As in the catabolism of OX-LDL. Moreover, the addition of GAGases to macrophages that were preincubated with anti-CD-36 antibodies and excess Ac-LDL further reduced macrophage degradation of Ox-LDL in comparison with cells that were pretreated only with anti-CD-36 antibodies and Ac-LDL, indicating a more complex role for endogenous GAGs. Overall, these studies demonstrate a substantial contribution of macrophage-associated GAGs in the catabolism of Ox-LDL, which is mediated in part by a cooperation between GAGs and cell-surface SR-As.

Beta-chemokines are released from HIV-1-specific cytolytic T-cell granules complexed to proteoglycans

CD8+ lymphocytes are believed to be important in host defence against the human immunodeficiency virus (HIV)-1, inhibiting HIV-1 replication through both cytolytic and non-cytolytic pathways. The cytolytic pathway involves calcium-dependent exocytosis of perforin and granzyme proteases, as well as Fas-mediated programmed cell death, whereas the noncytolytic pathway involves the release of chemokines that prevent viral entry. Using granzyme A as a marker of cytolytic granule proteins, and macrophage inflammatory protein (MIP)-1alpha and RANTES as markers of HIV-1 inhibitory chemokines, we show that these two very different mediators of viral inhibition are both localized in the cytolytic granules of HIV-1-specific CD8+ cytotoxic T lymphocytes (CTL). Following antigen-specific activation, these mediators are secreted together, facilitating both lysis of virion-producing cells and the inhibition of free virus. In addition, RANTES, MIP-1alpha and MIP-1beta are secreted by CTL as a macromolecular complex containing sulphated proteoglycans. This association appears to have a functional significance, because heparan sulphate facilitates RANTES inhibition of HIV-1 infection of monocytes.

Activated macrophages and the blood-brain barrier: inflammation after CNS injury leads to increases in putative inhibitory molecules

The cellular responses to spinal cord or brain injury include the production of molecules that modulate wound healing. This study examined the upregulation of chondroitin sulfate proteoglycans, a family of molecules present in the wound healing matrix that may inhibit axon regeneration in the central nervous system (CNS) after trauma. We have demonstrated increases in these putative inhibitory molecules in brain and spinal cord injury models, and we observed a close correlation between the tissue distribution of their upregulation and the presence of inflammation and a compromised blood-brain barrier. We determined that the presence of degenerating and dying axons injured by direct trauma does not provide a sufficient signal to induce the increases in proteoglycans observed after injury. Activated macrophages, their products, or other serum components that cross a compromised blood-brain barrier may provide a stimulus for changes in extracellular matrix molecules after CNS injury. While gliosis is associated with increased levels of proteoglycans, not all reactive astrocytes are associated with augmented amounts of these extracellular matrix molecules, which suggests a heterogeneity among glial cells that exhibit a reactive phenotype. Chondroitin sulfate also demarcates developing cavities of secondary necrosis, implicating these types of boundary molecules in the protective response of the CNS to trauma.

Macrophage-dependent regulation of syndecan gene expression

Heparan sulfates in the extracellular matrix are required for a variety of biological processes, including cellular response to heparin-binding growth factors. However, little is known regarding the regulation of their expression and composition under pathophysiological conditions. In the present study, we have investigated the regulation of expression of two key heparan sulfate chain-carrying core proteins, syndecan-1 and syndecan-4, in a mouse/rat infarct model of tissue injury and repair. Induction of myocardial infarction was associated with a prompt increase in expression of both syndecan genes. Although infiltrating macrophages accounted for a substantial increase in syndecan expression, increased expression was noted in the levels of syndecan-1 mRNA in endothelial cells and syndecan-4 mRNA in cardiac myocytes. This increase in expression was limited to the immediate peri-infarct region and was absent from remote areas of the left or right ventricles. The influx of blood-derived macrophages in the heart correlated with the appearance of PR-39 peptide, which has previously been shown to increase syndecan expression in vitro. Studies in the op/op mice strain (which demonstrates sharply reduced levels of circulating monocytes) showed that myocardial infarction was associated with markedly reduced levels of macrophage influx and corresponding reduction in the expression of PR-39 and both syndecan genes. Pretreatment of op/op mice with granulocyte macrophage colony-stimulating factor restored myocardial macrophage content with corresponding restoration of PR-39/syndecan expression. In summary, myocardial infarction is associated with a distinct spatial and temporal pattern of syndecan-1 and -4 gene expression, which is induced by an influx of blood-derived macrophages.

Proteoglycan synthesis in haematopoietic cells: isolation and characterization of heparan sulphate proteoglycans expressed by the bone-marrow stromal cell line MS-5

Proteoglycans of bone-marrow stromal cells and their extracellular matrix are important components of the haematopoietic microenvironment. Recently, several studies have indicated that they are involved in the interaction of haematopoietic stem and stromal cells. However, a detailed characterization of the heparan sulphate proteoglycans synthesized by bone-marrow stromal cells is still lacking. Here we report on the isolation and characterization of proteoglycans from the haematopoietic stromal cell line MS-5, that efficiently supports the growth and differentiation of human and murine haematopoietic progenitor cells. Biochemical characterization of purified proteoglycans revealed that the haematopoietic stromal cell line MS-5 synthesizes, in addition to chondroitin sulphate proteoglycans, several different heparan sulphate proteoglycans. Immunochemical analysis, using specific antibodies against the different members of the syndecan family, glypican, betaglycan and perlecan, showed that MS-5 cells synthesize all these different heparan sulphate proteoglycans. These data were further supported by reverse-transcriptase PCR and confirmed by sequence and Northern blot analysis. The relative abundance of the different heparan sulphate proteoglycans was estimated on the protein and mRNA levels.

Role of cell surface glycosaminoglycans of human T cells in human immunodeficiency virus type-1 (HIV-1) infection

To investigate the role of cell surface glycosaminoglycans (GAGs), including heparan sulfate (HS), on HIV-1 infection in human T cells, HIV-1 binding and infection were determined after treatment of T-cell lines and CD4+ T cells from normal peripheral blood mononuclear cells (PBMC) with GAG-degrading enzyme or a GAG metabolic sulfation inhibitor. Heparitinase I (hep I) and sodium chlorate prevented binding of HIV-1/IIIB to MT-4 cells as revealed by indirect immunofluorescence procedures, thereby inhibiting infection. Hep I was less effective in the binding inhibition of the macrophage-tropic strain HIV-1/SF162 than that of the T-cell line-tropic strain HIV-1/IIIB. The binding of HIV-1/SF162 was about 100-fold less dependent on cell surface HS than HIV-1/IIIB. Human HTLV-I positive T-cell lines expressed more HS than HTLV-I negative T-cell lines or normal CD4+ T cells when stained with anti-HS mAbs against either native or heparitinase-treated HS. With the exception of endo-beta-galactosidase (endo-beta-gal), GAG-degrading enzymes, including hep I, chondroitinase ABC (chon ABC), chondroitinase AC II (chon AC II) and keratanase, did not prevent the binding of HIV-1/IIIB to CD4+ T cells from normal PBMC. These results indicate that the cell surface HS of human T cells participates in HIV-1 infection by facilitating HIV-1/IIIB binding to MT-4 cells. In particular, the sulfation of HS chains is critical. Since the expression of cell surface HS varies among T cells, which are not consistently sensitive to hep I treatment in HIV-1 binding inhibition, other GAG-like molecules may also be involved.

Syndecan-1 alterations during the tumorigenic progression of human colonic Caco-2 cells induced by human Ha-ras or polyoma middle T oncogenes

The products of ras and src proto-oncogenes are frequently activated in a constitutive state in human colorectal cancer. In this study we attempted to establish whether the tumorigenic progression induced by oncogenic activation of p21ras and pp60c-src in human colonic Caco-2 cells is associated with specific alterations of syndecan-1, a membrane-anchored proteoglycan playing a role in cell-matrix interaction and neoplastic growth control. To this end, we used Caco-2 cells made highly tumorigenic by transfection with an activated (Val 12) human Ha-ras gene or with the polyoma middle T (Py-MT) oncogene, a constitutive activator of pp60c-src tyrosine kinase activity. Compared with control vector-transfected Caco-2 cells, both oncogene-transfected cell lines (1) contained smaller amounts of membrane-anchored PGs; (2) exhibited decreased syndecan-1 expression at the protein but not the mRNA level; (3) synthesized 35S-labelled syndecan-1 with decreased specific activity; (4) produced a syndecan-1 ectodomain with a lower molecular mass and reduced GAG chain size and sulphation; and (5) expressed heparanase degradative activity. These results show that the dramatic activation of the tumorigenic potential induced by oncogenic p21ras or Py-MT/pp60c-src in Caco-2 cells is associated with marked alterations of syndecan-1 expression at the translational and post-translational levels.

Proteoglycan synthesis in human and murine haematopoietic progenitor cell lines: isolation and characterization of a heparan sulphate proteoglycan as a major proteoglycan from the human haematopoietic cell line TF-1

Proteoglycans of bone-marrow stromal cells and their extracellular matrix are important components of the microenvironment of haematopoietic tissues. Proteoglycans might also be involved in the interaction of haematopoietic stem and stromal cells. Recently, several studies have been reported on the proteoglycan synthesis of stromal cells, but little is known about the proteoglycan synthesis of haematopoietic stem or progenitor cells. Here we report on the isolation and characterization of proteoglycans from two haematopoietic progenitor cell lines, the murine FDCP-Mix A4 and the human TF-1 cell line. Proteoglycans were isolated from metabolically labelled cells and purified by several chromatographic steps, including anion-exchange and size-exclusion chromatography. Biochemical characterization was performed by electrophoresis or gel-filtration chromatography before and after digestion with glycosaminoglycan-specific enzymes or HNO2 treatment. Whereas FDCP-Mix A4 cells synthesize a homogeneous chondroitin 4-sulphate proteoglycan, isolation and characterization of proteoglycans from the human cell line TF-1 revealed, that TF-1 cells synthesize, in addition to a chondroitin sulphate proteoglycan, a heparan sulphate proteoglycan as major proteoglycan. For this heparan sulphate proteoglycan a core protein size of approx. 59 kDa was determined. Immunochemical analysis of this heparan sulphate proteoglycan revealed that it is not related to the syndecan family nor to glypican.

Cell surface proteoglycans modulate net synthesis and secretion of macrophage apolipoprotein E

Using a macrophage cell line that constitutively expresses a human apolipoprotein E (apoE) cDNA, we have investigated the post-translational metabolism of endogenously produced apoE. Inhibition of lysosomal or cysteine proteases led to significant inhibition of apoE degradation but did not increase apoE secretion, indicating that cellular degradation is not limiting for apoE secretion in macrophages. Treatment of macrophages with inhibitors of proteoglycan synthesis (4-methylumbelliferyl-beta-D-xyloside) or sulfation (sodium chlorate) enhanced the release of apoE from cells and significantly attenuated the increase in secretion produced by incubation with phosphatidylcholine vesicles (PV). These observations suggested that a significant fraction of the apoE retained by cells (and released by incubation with PV) was associated with proteoglycans. Treatment of cells with exogenous heparinase led to a greater than 4-fold increase in apoE secretion and similarly attenuated the response to PV, suggesting that apoE was trapped in an extracellular proteoglycan matrix. This conclusion was confirmed in studies showing that PV could enhance the release of apoE from cells during an incubation at 4 degrees C, but this enhanced release was abolished in proteoglycan-depleted cells. Incubation with lactoferrin at 4 or 37 degrees C produced a similar decrement in cellular apoE, again indicating the existence of a cell surface pool of apoE. Pulse-chase studies showed that the apoE trapped in the proteoglycan matrix was susceptible to rapid cellular degradation such that net synthesis of apoE (secreted plus cell-associated) was increased significantly in proteoglycan-depleted cells compared with control cells as early as 45 min during a chase period.

Interactions between the Flk-1 receptor, vascular endothelial growth factor, and cell surface proteoglycan identified with a soluble receptor reagent

Fetal liver kinase-1 (Flk-1) is a transmembrane tyrosine kinase that was identified in endothelial cells and populations of cells enriched in hematopoietic progenitors. To characterize the interaction of Flk-1 with potential ligands the receptor extracellular domain was genetically fused to an alkaline phosphatase (AP) tag. A soluble ligand for Flk-1 was identified in the supernatants of numerous mesenchymal cell lines by co-immunoprecipitation with the Flk1-AP fusion protein. This polypeptide was shown by N-terminal sequencing to be vascular endothelial growth factor (VEGF). Receptor-AP fusion proteins can thus be used to identify soluble ligands as well as transmembrane ligands, and this approach is therefore likely to be widely applicable to many types of orphan receptor. The Flk1-AP soluble receptor was also found to bind to cell surfaces, showing two apparent classes of binding site with different affinities. This interaction could be reconstructed by introducing a VEGF expression plasmid into cells. These results indicate that VEGF presented at the cell surface can bind to the Flk-1 receptor, and could mediate a direct cell-cell interaction. The Flk1-AP fusion protein was also found to bind heparin, implying that ligand binding by the Flk-1 receptor may involve a three way interaction between the Flk-1 receptor, VEGF, and heparin-like cell surface proteoglycans.