- Email: [email protected]
signaling partnership of C1q receptors and integrins
International Immunopharmacology 3 (2003) 299 – 310 www.elsevier.com/locate/intimp
Complement component C1q induces endothelial cell adhesion and spreading through a docking/signaling partnership of $ C1q receptors and integrins Berhane Ghebrehiwet a,*, Xiaodong Feng b, Rajeev Kumar a, Ellinor I.B. Peerschke c a
Department of Medicine, Health Sciences Center, Division of Rheumatology, State University of New York, Stony Brook, NY 11794, USA b Department of Dermatology, State University of New York, Stony Brook, NY 11794, USA c Department of Pathology, Weill College of Medicine of Cornell University, New York, NY 10021, USA Accepted 17 September 2002
Abstract The interaction of C1q with endothelial cells elicits a multiplicity of biologic responses. Although these specific responses are thought to be mediated by the interaction of C1q with proteins of the endothelial cell surface, the molecular identity of the participant(s) has not been clearly defined. In this study, we examined the role of two C1q-binding proteins, cC1q-R/CR and gC1q-R/p33, on C1q-mediated adhesion and spreading of human dermal microvascular endothelial cells (HDMVECs). A specific and dose-dependent adhesion and spreading was observed when HDMVECs were cultured in microtiter plate wells coated with concentrations of C1q ranging from 0 to 50 Ag/ml. The extent of adhesion and spreading was similar to the adhesion seen on collagen-coated wells. Furthermore, the effect of C1q was mimicked by either polyclonal anti-cC1q-R or mAb 60.11, but not with isotype- and species-matched control IgG. More importantly, however, a 100% inhibition of spreading but not adhesion to C1q-coated wells was observed when HDMVECs were cultured in the presence of 30 mM of the peptide GRRGDSP but not GRRGESP. Furthermore, while anti-h1 integrin antibody blocked adhesion and spreading, antia5 integrin only blocked spreading. Since earlier studies have shown that zinc induces the exposure of hydrophobic sites in the C-terminus of gC1q-R including the putative high-molecular weight kininogen (HK)-binding site corresponding to residues 204 – 218, we also examined the effect of zinc on antibody binding to cell surface gC1q-R. Flow cytometric data show that the binding of mAb 74.5.2, which recognizes residues 204 – 218, is greatly enhanced when endothelial cells were incubated in the presence of 50 AM zinc. In summary, our data show that: (a) C1q-mediated endothelial cell adhesion and spreading requires the cooperation of both C1q receptors and 1 integrins, and possibly other membrane-spanning molecules, and (b) zinc can induce the exposure of hydrophobic sites in the C-terminal domain of gC1q-R allowing a more efficient binding of mAb 74.5.2 and HK. D 2002 Elsevier Science B.V. All rights reserved. Keywords: C1q; Endothelial cell; Kininogen
$ This paper is part of the Proceedings of the 16th International Conference on the Kallikrein-Kinin System (Kinin 2002) which was held in Charleston, SC, May 26 – 31, 2002 (see International Immunopharmacology Volume 2/13 – 14). * Corresponding author. Tel.: +1-631-444-2352; fax: +1-631-444-2493. E-mail address: [email protected] (B. Ghebrehiwet).
1567-5769/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. doi:10.1016/S1567-5769(02)00270-9
300
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
1. Introduction Surface properties of endothelial cells play a critical role in maintaining vessel wall integrity and in the pathophysiology of thrombosis, atherosclerosis and inflammation. This response is largely in part due to their ability to respond to a wide range of environmental signals with specific programmed responses (reviewed in Ref. [1]). Certain endothelial cell responses are rapid and do not require new gene transcription or protein synthesis. These include secretion of Weibel-Palade bodies with concomitant expression of surface P-selectin, activation of nitric oxide (NO) metabolic pathways and cytoskeletal or junctional rearrangement resulting in increased vascular permeability [1]. These responses are the hallmark of the early non-leukocyte-dependent phase of inflammation and a large number of agonists are capable of eliciting this type of response including proteins of the clotting and complement systems [1]. Various receptor systems have also been described on vascular cells that participate in the recognition, activation and clearance of components involved in humoral defense. Among these molecules are receptors for the collagen-like tail [cC1q-R or collectin receptor, a homologue of calreticulin (CR)] [2 – 4], and globular heads, gC1q-R/p33 (also known as p32) [5,6], of the complement component C1q. Interaction of C1q with endothelial cells in turn is known to induce a diversity of biological functions including adhesion and spreading [5]; stimulation and expression of the adhesion molecules E-selectin, ICAM-1 and VCAM-1 [7]; and production of IL-6, IL-8 and monocyte chemoattractant protein-1 [8]. Since C1q is present in significant quantities at sites of atherosclerotic and inflammatory and vascular lesions, and gC1q-R as well as cC1q-R are present on endothelial cells, modulation of endothelial cell function by soluble and immobilized C1q may contribute significantly to the development of thrombosis and inflammation. In addition, endothelial cell gC1q-R [9,10], together with the urokinase plasminogen activator receptor (u-PAR) [11], and cytokeratin 1 [12], has been shown to serve as a zinc-dependent, high-affinity site for high-molecular weight kininogen (HK) and factor XII (FXII) [9] that leads to the generation of the potent multifunctional peptide, bradykinin [13]. The present studies were therefore undertaken to address
the following specific questions: (a) which C1q receptors and/or other molecules are involved in the C1q-mediated activation of endothelial cells? (b) Does zinc expose neoepitopes on endothelial cell surface so that HK-binding to gC1q-R is enhanced?
2. Materials and methods 2.1. Chemicals and reagents Unless specified elsewhere, the following chemicals and reagents were purchased from the commercial sources indicated: FCS (Hyclone Laboratories, Logan, UT): RPMI 1640, 100 antibiotic antimycotic mixture, Dulbecco’s phosphate buffered saline (D-PBS), GRRGDSP and GRRGESP peptides, as well as anti-a5 integrin (P1D6) (GIBCO BRL, Gaithersburg, MD); mAb (4B4) antibody to h1 integrin (Coulter Immunology, Hialeah, FL). 2.2. Proteins Highly purified human C1q, free of the C1q inhibitor chondroitin 4-sulfate proteoglycan [14], was either purified as described [15] or purchased from Advanced Research Technologies (San Diego, CA). Before use in culture with endothelial cells, the C1q was dialyzed against sterile RPMI or D-PBS before use in culture with endothelial cells. Heatinactivated C1q was prepared by incubating C1q (1 mg/ml) at 56 jC for 1 h. The residual activity of inactivated C1q (DC1q) was monitored by comparing its activity with untreated C1q by a standard hemolytic assay using C1q-depleted (C1qD) serum as described earlier [16]. 2.3. Antibodies The production and characterization of monoclonal antibodies (mAb) 60.11 and 74.5.2 as well as polyclonal antibodies (pAb) to gC1q-R and to cC1q-R in rabbits have been described earlier [17]. The mAb 60.11 is directed against the N-terminal segment containing a C1q-binding site, whereas the 74.5.2 recognizes epitopes in a segment of gC1q-R containing a site for high-molecular weight kininogen. The IgG fraction was purified as described in detail
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
previously [18]. Monoclonal antibody specific for h1 integrin (mAb 4B4) was purchased from Coulter Immunology and anti-a5 integrin (mAb P1D6) was purchased from GIBCO BRL. The following monoclonal antibodies: A3D8 (anti-CD 44, Mr 80 – 110 kDa), AF3 (anti-CD44H, Mr 100 kDa), P1B5 (antia3 integrin), a4 (anti-a4 subunit of VLA 4) and P1D4 (anti-a5 subunit of integrin) were a generous gift from Dr. Michael Shepley, Division of Infectious Diseases, SUNY at Stony Brook. 2.4. Culture of human dermal microvascular endothelial cells (HDMVECs) Human dermal microvascular endothelial cells were isolated from human neonatal foreskins as previously reported [19]. Briefly, after initial harvest from minced trypsinized human foreskins, microvascular endothelial cells were further purified on a Percoll density gradient. HDMVECs were cultured on collagen type 1 coated tissue culture flasks in endothelial cell growth medium (EGM) consisting of endothelial basal medium (EBM) supplemented with 10 ng/ml epidermal growth factor, 0.4% bovine brain extract, 17.5 Ag/ml dibutyryl cyclic AMP and 1 Ag/ml hydrocortisone in the presence of 30% normal human serum. Endothelial cell cultures were characterized and determined to be >99% pure on the basis of formation of typical cobblestone monolayers in culture, positive immunostaining for von Willebrand factor and selective uptake of acetylated low-density lipoprotein. All experiments were done with HDMVECs below passage 10 as described in our recent paper [20]. 2.5. Endothelial cell adhesion and spreading assay The assay used here for measuring endothelial cell adhesion and spreading was modified from a previously described standard assay [21]. Immulon-4 microtiter plates (Dynatech Laboratories, Chantilly, VA) were coated with 50 Al/well of C1q (10 –50 Ag/ ml) or other purified proteins diluted in 50 mM Tris pH 7.4, 150 mM NaCl and 0.1% NaN3. Bovine serum albumin (BSA, 20 mg/ml) or DC1q were used as negative controls, while type I collagen (10 Ag/ml) was used as positive control. Plates were incubated overnight at 4 jC. All plates were then washed,
301
incubated with 2% BSA in PBS containing 0.1% NaN3 for 2 h at room temperature (20 jC) to block nonspecific binding sites, and then washed again prior to use. HDMVECs were washed twice, harvested, resuspended in assay buffer (EBM with 0.1% BSA) and incubated with or without inhibitors for 30 min at room temperature prior to initiation of the assay. Cell attachment and spreading was measured by adding 100 Al aliquots of HDMVEC suspension (104 cells/ well) to the coated microtiter plate wells and incubating the plate at 37 jC for 1 – 2 h for attachment and 6– 8 h for spreading. To assess the potential effect of antibodies on cell adhesion or spreading, the cells were pre-treated (30 min, 37 jC) with a predetermined concentration of the anti-C1q-R or anti-integrin antibodies before addition to the C1q-coated plates. At the end of the assay, 100 Al of 2% glutaraldehyde was carefully added to each well to fix the attached cells. Then, the unattached cells were removed by gently washing the wells two times with PBS and one time with distilled water. Endothelial cell adhesion and spreading was observed and recorded with a Nikon Diaphot-TMD inverted microscope (Nikon, Melville, NY), equipped with a video system consisting of a Dage-MTI CCD-72S video camera and linked to a Macintosh G3 computer. The images were captured at various magnifications using Adobe Photoshop. All experiments were repeated at least three times. 2.6. Quantification of cell adherence and spreading The percent of adherent cells was quantified spectrophotometrically by the detergent-compatible bicinchoninic acid method after solubilizing the cells with 4,4-dicarboxy-2,2V-biquinoline (Pierce). Nonspecific adhesion to BSA- or DC1q-coated wells was subtracted from each experiment. Inhibition of spreading was calculated by subtracting the number of spreading cells counted under the microscope from the total adherent cells in each experiment as described earlier [20]. 2.7. Cell surface biotinylation of HDMVECs The confluent HDMVECs monolayers (106/ml) grown in a collagen-coated flask were surface biotinylated in situ using the surface impermeable reagent
302
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
sulfo-NHS-LC-biotin as described previously [18,20]. The adherent cells in the flask were washed three times in warm HEPES-buffered saline (HBS) (10 mM HEPES, 137 mM NaCl, 4 mM KCl, 11 mM glucose) and incubated (4 jC, 2 h) with 10 ml of 5 mM sulfoNHS-LC-biotin. After incubation, the excess biotin was removed; the cells were washed twice in 20 ml HBS and lysed using lysis buffer [10 mM HEPES, 150 mM NaCl, 2 mM phenylmethylsulfonyl fluoride (PMSF), 1 AM aprotinin, 1 AM pepstatin, 1 mM EDTA and 0.1% soybean trypsin inhibitor (SBTI) and 1% NP-40] on ice. The cell lysate was finally taken out from the flask, transferred onto sterile test tubes and, after removal of the nuclei and insoluble cellular debris by centrifugation (15 min, 850 g, 4 jC), the supernatant containing the labeled proteins was subjected to further centrifugation (1 h, 45,000 g, 4 jC). The supernatant was then collected, total protein concentration estimated by the BCA protein assay (Pierce) and the degree of biotinylation verified by ELISA using alkaline phosphatase-conjugated streptavidin as a probe. The labeled protein solution was used
either immediately or aliquoted and kept frozen at 80 jC. 2.8. Antigen capture enzyme-linked immunosorbent assay (AC-ELISA) For the AC-ELISA, microtiter plates (MaxiSorb, Nunc, Denmark) were first coated with 100 Al of a 10 Ag/ml (carbonate buffer pH 9.5) of either the capturing polyclonal anti-gC1q-R antibody or non-immune, species-matched control IgG (2 h, 37 jC), washed with TBS-T, blocked with 2% BSA and, after further washing, 100 Al of biotinylated HDMVEC membranes that had been diluted to a concentration of 100 Ag/ml in TB containing 0.5 M NaCl, 0.05% Tween 20 and 0.1% BSA, were added to each well and incubated (overnight, 4 jC). After incubation, the wells were washed once with TB containing 1 M NaCl, 0.05% Tween 20, twice with TBST and once with TBS. Washes in high salt were essential in order to reduce non-specific binding. The captured proteins were then detected using monoclonal antibodies to
Fig. 1. Dose-dependent adhesion of endothelial cells to C1q. Endothelial cells (1 105/well) were incubated (overnight, 4 jC) in 24-well microtiter plate wells coated with either 1 mg/ml BSA, or concentrations of C1q ranging from 10 to 50 Ag/ml. After incubation, the cells were washed three times with PBS, and the adherent cells were fixed and visualized under the microscope. The figure is a representative of four experiments run in duplicates (adapted from Ref. [20]).
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310 Table 1 Effect of anti C1q-R antibodies on endothelial cell spreading and/or adhesion Antibody (30 Ag/ml)
Spreading (% F SDE)
Adhesion (% F SDE)
Anti-gC1q-R (mAb 60.11) Anti-cC1q-R (pAb)
68 F 12 54 F 7
47 F 9 39 F 5
Cells were first pre-treated (30 min, 4 jC) with either 30 Ag/ml of control rabbit IgG, pAb anti-cC1q-R (polyclonal antibody) or mAb 60.11 before addition to C1q-coated wells. The data represent values after subtraction of non-specific adhesion or spreading with IgG controls.
either gC1q-R or integrins and further developed by standard ELISA. Standard for the capture assay included concentrations (0– 1000 ng/ml, in TBS containing 0.5 M NaCl) of highly purified gC1q-R, whereas a similarly treated irrelevant antigen, BSA, was used as control for non-specificity. To ensure that the captured antigens are surface-labeled, a duplicate AC-ELISA was performed under the same conditions except that the captured antigens were detected by alkaline phosphatase-conjugated streptavidin or extravidin and visualized by reaction with p-nitrophenyl phosphate (pNPP).
303
ISS PC1 photon counting spectrophotometer at 25 jC. Each sample contained 0.1 Amol of gC1q-R and 1 Amol of bis-ANS either in HEPES only or with one of the salts at a final concentration of 500 AM: KCl, CaCl2, MnCl2, MgCl2, ZnCl2 and ZnCl2 +EDTA. 2.10. Effect of zinc on endothelial cells To investigate the effect of zinc on surface expressed gC1q-R, flow cytometric analysis was performed using human bone marrow endothelial cells (HBMECs). HBMECs were first pre-incubated with 5 Ag/ml human Fc fragments to block Fc receptors followed by incubation (1 h, 37 jC) with either mAb 74.5.2 alone or in the presence of 50 AM zinc. Isotype- and species-matched non-immune IgG instead of mAb was used as control for non-specificity. After incubation, the cells were washed and the bound antibody probed with Alexa 488-conjugated F(abV)2 goat anti-mouse antibody.
3. Results
2.9. Fluorescence spectroscopy
3.1. Adhesion and spreading of endothelial cells on C1q
Bis-ANS fluorescence experiments were performed with an excitation wavelength at 395 nm and emission wavelength from 430 to 560 nm on
Previous studies have shown that various types of endothelial cells are capable of binding human C1q and that this binding can trigger biological responses
Fig. 2. RGD inhibits C1q-mediated endothelial cell spreading. Cells were incubated (30 min, 4 jC) with either 30 AM RGE or RGD before addition to C1q-coated wells. After incubation, the cells were treated and processed as above. The figure is a representative of four such experiments run in duplicates (adapted from Ref. [20]).
304
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
that include adhesion and spreading. Employing cultured HDMVECs and C1q-coated microplate wells as a model system, we show here that these cells also are able to specifically bind to and spread on C1q (Fig. 1). The binding and spreading was dose-dependent and at physiologic ionic strength, and did not require the presence of metal ions. Furthermore, adhesion to C1q was inhibited when the binding was performed in the presence of (FabV)2 anti-C1q antibody or when ?C1q was used instead of C1q (not shown). No adhesion or spreading occurred on wells that had been coated with 1 mg/ml BSA. The adhesion of HDMVECs to C1q was found to be quantitatively and qualitatively similar to that of collagen type 1 (Fig. 1). 3.2. Spreading of HDMVECs to C1q is inhibited by anti-gC1q-R and anti-cC1q-R To investigate whether the endothelial cell binding is mediated by cC1q-R and/or gC1q-R, HDMVECs were first pre-incubated (30 min, 4 jC) with concentrations ranging from 10 to 100 Ag/ml of either control IgG or mAb 60.11 before addition to C1q-coated wells. After incubation, the cells were washed and the adherent cells fixed and visualized as described. Table 1 shows that spreading (68% F 12) and, to a moderate degree (47% F 9), adhesion was inhibited by mAb 60.11. This inhibition was dose-dependent but no inhibition was noted with even the highest dose (100 Ag/ml) of control IgG. Similar results were obtained when endothelial cells were pre-incubated with 30 Ag/ml pAb anti-cC1q-R antibody (Table 1). When the two antibodies were mixed in the proportion of 100 (60.11) and 30 Ag/ml (pAb anti-cC1q-R), they had a slight (75% F 13) additive effect in their ability to inhibit spreading. 3.3. RGD inhibits C1q-mediated endothelial cell spreading Using human diploid fibroblasts, it has been shown previously that adhesion of cells to C1q was inhibited by soluble GRGDTP peptide [22]. This suggested that adhesion of cells to C1q may require the participation of C1q receptors and integrins. To test this hypothesis, HDMVECs were incubated (30 min, 4 jC) with either 30 AM GRRGDSP (RGD) or GRRGESP (RGE)
Fig. 3. Effect of monoclonal antibodies to h1 and a5 integrins on endothelial cell spreading. Endothelial cells were first pretreated (30 min, 4 jC) with either 50 Ag/ml control IgG or mAb anti-b1 integrin. The cells (1 105/well) were then added to wells coated with C1q, and visually examined for spreading and adhesion as described above.
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
305
before addition to C1q-coated wells. After incubation, cells were treated and processed as described above. As shown in Fig. 2, RGD peptide but not RGE was able to completely (100%) inhibit endothelial cell spreading but not adhesion as assessed by visual examination and manual counting. Those that remained adherent were round and without any of the characteristic cytoskeletal reorganization and formation of typical cobblestone monolayers seen on spreading cells.
sion as described above. Fig. 3 is a representative experiment and shows that, while mAb anti-h1 integrin inhibited both spreading and adhesion of HDMVECs to C1q-coated wells, anti-a5 integrin (not shown) was able to inhibit only cell spreading. Additional experiments, which were similarly designed to test the effect of other integrins, had minimal or no significant inhibitory effect.
3.4. Effect of monoclonal antibody to b1 integrin on endothelial cell spreading
The fluorescence intensity of bis-ANS greatly increases upon binding to hydrophobic sites and has been used widely to assess the hydrophobic surface of proteins [23]. A blue shift in emission maximum is also observed upon binding of bis-ANS to hydrophobic surface in comparison to emission maximum of 520 nm for free bis-ANS. Fig. 4 shows cationinduced change of bis-ANS fluorescence by gC1q-R changing the bis-ANS emission spectra from 520 to 485 nm. Addition of zinc further blue-shifted the bisANS emission spectra from 485 to 475 nm and produced three- to four-fold enhancement in fluorescence intensity. Addition of EDTA to the solution containing gC1q-R, bis-ANS and zinc results in a four-fold decrease in the bis-ANS fluorescence intensity indicating that the zinc-induced hydrophobic exposure is reversible. Furthermore, it was demon-
To identify the type of integrin(s) that may be involved in C1q-mediated endothelial cell spreading, AC-ELISA was first performed on HDMVEC membrane proteins using pAb anti-gC1q-R-coated wells. The captured proteins were then probed with various anti-integrin antibodies as described in Section 2. The results of this experiment showed that h1 integrin is cocaptured (not shown) with gC1q-R giving strong evidence that h1 integrins and gC1q-R may collaborate to induce C1q-mediated spreading. To test this hypothesis, HDMVECs were first pre-treated (30 min, 4 jC) with either 50 Ag/ml control IgG or mAb anti-h1. The cells (1 105/well) were then added to C1q-coated wells and visually examined for spreading and adhe-
3.5. Zinc can induce hydrophobic exposure in gC1q-R
Fig. 4. Bis-ANS fluorescence emission spectra. Emission spectra from bottom to top are: (i) ZnCl2 + EDTA, CaCl2, buffer alone, ZnCl2.
306
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
Fig. 5. Flow cytometric analysis of endothelial cells. Endothelial cells were reacted with mAb 74.5.2 in the presence [3] or absence [2] of 50 AM ZnCl2. The bound antibody was probed with Alexa-488conjugated F(abV)2 goat anti-mouse antibody. Species- and isotypematched IgG (MOPC) was used as control for non-specificity [1]. The figure is a representative of five experiments.
strated that the zinc-induced exposure of hydrophobic surface is not the result of increased nonspecific ionic strength, since the fluorescence intensity of bisANS + gC1q-R did not increase in the presence of other divalent or monovalent cations, such as MgCl2, MnCl2, CaCl2 and KCl. 3.6. Endothelial cells express enhanced antibody binding in the presence of zinc Endothelial cells express gC1q-R on their surface as evidenced by moderate binding of either mAb 60.11 or 74.5.2. However, while the binding of mAb 74.5.2 to endothelial cells is enhanced in the presence of zinc (Fig. 5), no increase in binding was noted with mAb 60.11 (not shown), which recognizes an epitope in the N-terminal region of gC1q-R under the same experimental conditions. No significant staining was also observed with either control IgG alone or in the presence of zinc (not shown). Similar results were also observed with platelets and the cell line U937.
4. Discussion The ability of C1q to interact with various types of vascular endothelial cells leading to the production of biologically active proteins or expression of adhesive
molecules has been well documented. Previously, we and others have shown that endothelial cells constitutively express both cC1q-R/CR and gC1q-R [3– 5] and that the expression of these molecules is upregulated by inflammatory cytokines [6]. However, despite the availability of data implicating these molecules to play a significant role in various C1q-mediated endothelial cell function, the lack of a ‘‘consensus motif’’ in their respective sequences consistent with a transmembrane domain has made it difficult to explain how these molecules communicate with elements inside the cell. The present studies were, therefore, undertaken to explore the possibility that such molecules may circumvent their lack of direct access to the interior of the cell by forging an association with membrane-spanning cell surface proteins in a manner similar to uPAR [24 – 26]. To address this question, we employed an experimental system in which the adherence and spreading of human dermal microvascular endothelial cells on C1q-coated microtiter wells was investigated, and the involvement of the two C1q receptors in question, cC1q-R/CR and or gC1q-R, was assessed. Our results show that HDMVECs, like other endothelial cell types, can adhere and spread on C1q-coated microtiter wells in a specific and dose-dependent fashion (Fig. 1), and this finding is consistent with the long-held postulate that C1q can act as a matrix protein for endothelial cells and fibroblasts [4,5,22]. The degree of adherence and spreading was qualitatively and quantitatively comparable to the adherence of HDMVECs on type I collagen. Furthermore, while adherence was not drastically affected, spreading of HDMVEC to C1q was inhibited when the cells were first pre-treated with either mAb 60.11, an antibody directed to a C1q-binding site on the N-terminus of gC1q-R or polyclonal anti-cC1q-R antibodies, while species- and isotype-matched control IgG had no effect on either adherence or spreading. More importantly, however, spreading was completely and adherence moderately inhibited when HDMVECs were preincubated with 30 AM of soluble RGD peptide but not RGE before addition to C1q-coated wells (Fig. 2). Moreover, pre-treatment of HDMVECs with 50 Ag/ml of mAb anti-h1 integrin before addition to C1q-coated wells also resulted in the inhibition of adhesion and spreading, but pre-incubation of cells with control IgG did not. Of the anti-a integrins tested, only anti-a5
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
integrin was able to moderately inhibit spreading indicating that it may be a likely partner of the h1 integrin. AC-ELISA using anti-gC1q-R also confirmed that h1 integrin is co-captured from solubilized HDMVEC membranes. The studies were undertaken to answer a simple but intriguing biological question: how do proteins like cC1q-R and gC1q R, which lack a traditional membrane spanning domain, communicate with elements inside the cell? The results presented in this report collectively suggest that, at least on HDMVECs, C1q-mediated adhesion and spreading may require the participation of cell surface C1q receptors and h1 integrins, and possibly other molecules. This model would envisage that both cC1q-R and gC1q-R, which form a high-affinity binding complex upon ligand binding [18], would laterally associate with h1 integrin to form a docking/signaling complex. This kind of molecular association between membrane-spanning and non-membrane-spanning proteins to form a docking/signaling complex is not unique to gC1q-R and/or cC1q-R. Many surfaceassociated proteins, which lack transmembrane domains, are known to trigger biological responses using this strategy. For example, urokinase plasminogen activator receptor (CD87), which is also expressed on vascular endothelial cells, is a GPIanchored protein and, as such, has no direct link with signaling proteins inside the cell [24]. However, uPAR can form a complex with h1 or h2 integrins to modulate their adhesive functions and with the h2 integrin, CR3 (CD11b/CD18), to trigger uPA-induced Ca2 + fluxes in neutrophils [24,25]. More recently, a signaling partnership between uPAR and L-selectin (CD62L) was also demonstrated in human polymorphonuclear neutrophils [26]. These findings, therefore, would tend to support the concept that membrane-associated proteins can pick and choose their signaling partners depending on the cell type and signal required. Recent evidence also suggests that C1q and mannose-binding lectin (MBL) can engage cell surface calreticulin (cC1q-R) and CD91 to initiate macropinocytosis and uptake of apoptotic cells. MBL is a member of the family of proteins collectively known as collectins (collagen containing lectins) and, like C1q, binds to cC1q-R or calreticulin (or collectin receptor) [27,28]. Apoptotic cell uptake via the calreticulin/CD91 docking/signaling complex is, therefore,
307
another example of the molecular partnership that is forged between transmembrane and non-transmembrane proteins in the induction of specific biological responses. In addition to its participation in C1q-mediated responses, gC1q-R has been shown to serve as a high-affinity receptor for high-molecular weight kininogen. The binding of HK to gC1q-R is entirely dependent on the presence of zinc [9,10] and is inhibited by mAb 74.5.2, which recognizes an epitope within the gC1q-R segment corresponding to residues 204– 218. This segment, which forms part of the h6 strand in the crystal structure [29], is largely restricted to the inside of the central hole (Fig. 6). The homotrimeric, doughnut-like gC1q-R molecule has two arbitrarily assigned faces [30]. This assignment is primarily based on the large difference in charge between the two faces. One side is highly acidic and is likely going to interact with ligands in solution (Sface), whereas the mildly basic face interacts with the cell membrane (M-face). Although mAb 74.5.2 can recognize gC1q-R on the quiescent endothelial cell surface, its interaction is greatly enhanced in the presence of 50 AM zinc. This is largely because zinc can induce exposure of hydrophobic sites in the Cterminal domain of gC1q-R, which includes residues 204– 218 (Figs. 3 and 4). It is under these conditions we postulate that the binding of HK to the endothelial cell gC1q-R becomes greatly increased and this highaffinity (9 F 2 AM) interaction would then trigger the activation of the kinin-forming cascade. Bradykinin, thus generated, would in turn bind to the endothelial cell bradykinin receptor (B2) and induce a plethora of pathophysiologic responses ranging from tumor metastasis and angiogenesis to the induction of morphologic changes in the endothelium rendering the subendothelial matrix accessible to blood components, thus leading to infiltration by inflammatory cells. Although both cC1q-R and gC1q-R are constitutively expressed on resting nonthrombotic endothelium and gC1q-R in particular has the potential to activate the bradykinin-generating system [13], the mechanism by which a continuously thrombogenic state of the endothelium is averted is not known. Based on the above findings, we speculate that efficient engagement of gC1 q-R by its proinflammatory ligands or C1q-containing immune complexes is re-
308
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
Fig. 6. The localization of antibody binding sites. The localization of mAb 60.11, which recognizes epitopes in the N-terminus containing a C1q-binding site is predominantly exposed on the S face of the molecule, whereas mAb 74.5.2 recognizing the HK site is restricted to the inside of the hole of the homotrimeric molecule.
stricted to conditions where, under chemical, physical or infectious insult, the endothelial cell is converted to a prothrombotic and proinflammatory phenotype leading to the induction of cytokines, such as IL-1 or TNFa or expression of cell adhesion molecules (CAMs) [1]. Leukocytes bound to the cell adhesion molecules can release cytokines, which in turn can amplify the system by upregulating the expression of C1q-binding molecules in a manner that allows efficient binding of ligands such as C1q or HK. That cytokines, such as LPS, IFNg and TNFa, can upregulate the expression of both cC1q-R and gC1q-R has been reported previously [6]. Thus, modulation of
endothelial cell function by soluble and/or immobilized proinflammatory ligands such as C1q or HK especially at sites of inflammatory and vascular lesions may contribute significantly to the development of thrombosis and exacerbation of the inflammatory process.
Acknowledgements This work was supported in part by grants RPG95068-06 from the American Cancer Society (B.G.), R01 HL5029101 from National Heart Blood and
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
Lung Institute (E.I.B.P. and B.G.) and a generous gift from Larry and Sheila Dalzell (B.G.). We are indebted to the late Dr. Michael Shepley, Division of Infectious Diseases, SUNY Stony Brook, without whose generous gift of the anti-integrin antibodies, this work would not have been completed. The expert technical assistance of Dana Jaggerssarsingh and Lynda Piboon is also greatly acknowledged.
References [1] Silverstein RL. The vascular endothelium. In: Gallin JI, Snyderman R, editors. Inflammation: Basic Principles and Clinical Correlates. Philadelphia: Lippincott Williams and Wilkins; 1999. p. 207. [2] Andrews BS, Shadforth M, Cunningham P, Davis IV JS. Demonstration of a C1q receptor on the surface of human endothelial cells. J Immunol 1981;127:1075. [3] Peerschke EIB, Malhotra R, Ghebrehiwet B, Reid KBM, Willis AC, Sim RB. Isolation of a human endothelial cell C1q receptor (C1q-R). J Leukoc Biol 1993;53:179. [4] Daha MR, Dunn L, Van den Berg RH, Muizert-de Lange Y, Gerritsen AF, Van Eys LA. Isolation of a human endothelial cell C1q receptor. Mediat Inflamm 1993;2:447. [5] Peerschke EIB, Smyth SS, Teng E, Dalzell M, Ghebrehiwet B. Human umbilical vein endothelial cells possess binding sites for the globular domains of C1q. J Immunol 1996;157:4154. [6] Guo W-X, Ghebrehiwet B, Weksler B, Schweizer K, Peerschke EIB. Upregulation of endothelial cell binding proteins/ receptors for complement C1q by inflammatory cytokines. J Lab Clin Med 1999;133:541. [7] Lozada C, Levin RI, Huie M, Hirschhorn R, Naime D, Whitlow M, et al. Identification of C1q as a heat-labile serum cofactor required for immune complexes to stimulate endothelial expression of the adhesion molecules E-selectin and intercellular and vascular cell adhesion molecules 1. Proc Natl Acad Sci U S A 1995;92:8378. [8] van den Berg RH, Faber-Krol MC, Sim RB, Daha MR. The first subcomponent of complement, C1q, triggers the production of IL-8, IL-6 and monocyte chemoattractant peptide-1 by human umbilical vein endothelial cells. J Immunol 1998; 161:6924. [9] Joseph K, Ghebrehiwet B, Peerschke EIB, Reid KBM, Kaplan AP. Identification of the zinc-dependent endothelial cell binding protein for high molecular weight kininogen and factor XII: identity with the receptor which binds to the globular ‘heads’ of C1q (gC1q-R). Proc Natl Acad Sci U S A 1996; 93:8552. [10] Herwald H, Dedio J, Kellner R, Loos M, Muller-Esterl W. Isolation and characterization of the kininogen-binding protein p33 from endothelial cells. Identity with the gC1q receptor. J Biol Chem 1996;271:13040.
309
[11] Coleman RW, Pixley RA, Najamunnisa S, Yan W, Wang J, Mazer A, et al. Binding of high molecular kininogen to human endothelial cells is mediated via a site within domains 2 and 3 of the urokinase receptor. J Clin Invest 1997;100:1481. [12] Hasan AAK, Zisman T, Schmaier AH. Identification of cytokeratin 1 as a binding protein and presentation receptor for kininogens on endothelial cells. Proc Natl Acad Sci U S A 1998;95:3615. [13] Joseph K, Ghebrehiwet B, Kaplan AP. Activation of the kininforming cascade on the surface of endothelial cells. Biol Chem 2001;382:71. [14] Kolb WP, Kolb LM, Podack ER. C1q: isolation from human serum in high yield by affinity chromatography and development of a highly sensitive hemolytic assay. J Immunol 1979; 122:2103. [15] Silvestri L, Baker JR, Roden L, Stroud RM. The C1q inhibitor in serum is a chondroitin 4-sulfate proteoglycan. J Biol Chem 1981;256:7383 – 7. [16] Conradie JD, Volanakis JE, Stroud RM. Evidence for a serum inhibitor of C1q. Immunochemistry 1975;12:967 – 71. [17] Ghebrehiwet B, Lu PD, Zhang W, Lim B-L, Eggleton P, Leigh LEA, et al. Identification of functional domains on gC1q-R, a cell surface protein which binds to the globular ‘‘heads’’ of C1q, using monoclonal antibodies and synthetic peptides. Hybridoma 1996;15:333. [18] Ghebrehiwet B, Lu PD, Zhang W, Keilbaugh SA, Leigh LEA, Eggleton P, et al. Evidence that the two C1q-binding membrane proteins gC1q-R and cC1q R associate to form a metal ion-independent complex. J Immunol 1997;159:1429. [19] Feng X, Clark RAF, Galanakis D, Tonnesen MG. Fibrin and collagen differentially regulate human dermal microvascular endothelial cell integrins: stabilization of av/b3 mRNA by fibrin. J Invest Dermatol 1999;113:913. [20] Feng X, Tonnesen MG, Peerschke EIB, Ghebrehiwet B. Cooperation of C1q receptors and integrins in C1q-mediated endothelial cell adhesion and spreading. J Immunol 2002; 168:2441 – 8. [21] Gailit J, Clark RAF. Studies in vitro on the role of av and h1 integrins in the adhesion of human dermal fibroblasts to provisional matrix proteins fibronectin, vitronectin, and fibrinogen. J Invest Dermatol 1996;106:102. [22] Bordin S, Ghebrehiwet B, Page RC. Participation of C1q and its receptor in adherence of human diploid fibroblast. J Immunol 1990;145:2520. [23] Cardamone M, Puri NK. Spectrofluorimetric assessment of the surface hydrophobicity of proteins. Biochem J 1992;282: 589 – 93. [24] Sitrin RG, Pan PM, Harper HA, Blackwood RA, Todd III RF. Urokinase receptor (CD87) aggregation triggers phosphoinositide and intracellular calcium mobilization in mononuclear phagocytes. J Immunol 1999;163:6193. [25] Sitrin RG, Todd III RF, Petty HR, Brock TG, Shollenberger SB, Abrecht E, et al. The urokinase receptor (CD87) facilitates CD11b/CD18-mediated adhesion of human monocytes. J Clin Invest 1996;97:1942. [26] Sitrin RG, Pan PM, Blackwood RA, Huang J, Petty HR. Evidence for signaling partnership between urokinase recep-
310
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
tors (CD87) and L-selectin (CD62L) in human polymorphonuclear neutrophils. J Immunol 2001;166:4822. [27] Sim RB, Moestrup SK, Stuart GR, Lynch NJ, Lu J, Schwaeble WJ, et al. Interaction of C1q and the collectins with the potential receptors calreticulin (cC1qR/collectin receptor) and megalin. Immunobiology 1998;199:208. [28] Malhotra R, Lu J, Thiel S, Jensenius J-C, Willis AC, Sim RB. C1q receptor (collectin receptor): primary structure, ho-
mology, and interaction with ligands. Immunobiology 1992; 184:1437. [29] Jiang J, Zhang Y, Krainer AR, Xu RM. Crystal structure of human p32, a doughnut-shaped acidic mitochondrial matrix protein. Proc Natl Acad Sci U S A 1999;96:3572 – 7. [30] Ghebrehiwet B, Jesty J, Peerschke EIB. gC1q-R/p33: structure – function predictions from the crystal structure. Immunobiology 2002;205:421.
Complement component C1q induces endothelial cell adhesion and spreading through a docking/signaling partnership of $ C1q receptors and integrins Berhane Ghebrehiwet a,*, Xiaodong Feng b, Rajeev Kumar a, Ellinor I.B. Peerschke c a
Department of Medicine, Health Sciences Center, Division of Rheumatology, State University of New York, Stony Brook, NY 11794, USA b Department of Dermatology, State University of New York, Stony Brook, NY 11794, USA c Department of Pathology, Weill College of Medicine of Cornell University, New York, NY 10021, USA Accepted 17 September 2002
Abstract The interaction of C1q with endothelial cells elicits a multiplicity of biologic responses. Although these specific responses are thought to be mediated by the interaction of C1q with proteins of the endothelial cell surface, the molecular identity of the participant(s) has not been clearly defined. In this study, we examined the role of two C1q-binding proteins, cC1q-R/CR and gC1q-R/p33, on C1q-mediated adhesion and spreading of human dermal microvascular endothelial cells (HDMVECs). A specific and dose-dependent adhesion and spreading was observed when HDMVECs were cultured in microtiter plate wells coated with concentrations of C1q ranging from 0 to 50 Ag/ml. The extent of adhesion and spreading was similar to the adhesion seen on collagen-coated wells. Furthermore, the effect of C1q was mimicked by either polyclonal anti-cC1q-R or mAb 60.11, but not with isotype- and species-matched control IgG. More importantly, however, a 100% inhibition of spreading but not adhesion to C1q-coated wells was observed when HDMVECs were cultured in the presence of 30 mM of the peptide GRRGDSP but not GRRGESP. Furthermore, while anti-h1 integrin antibody blocked adhesion and spreading, antia5 integrin only blocked spreading. Since earlier studies have shown that zinc induces the exposure of hydrophobic sites in the C-terminus of gC1q-R including the putative high-molecular weight kininogen (HK)-binding site corresponding to residues 204 – 218, we also examined the effect of zinc on antibody binding to cell surface gC1q-R. Flow cytometric data show that the binding of mAb 74.5.2, which recognizes residues 204 – 218, is greatly enhanced when endothelial cells were incubated in the presence of 50 AM zinc. In summary, our data show that: (a) C1q-mediated endothelial cell adhesion and spreading requires the cooperation of both C1q receptors and 1 integrins, and possibly other membrane-spanning molecules, and (b) zinc can induce the exposure of hydrophobic sites in the C-terminal domain of gC1q-R allowing a more efficient binding of mAb 74.5.2 and HK. D 2002 Elsevier Science B.V. All rights reserved. Keywords: C1q; Endothelial cell; Kininogen
$ This paper is part of the Proceedings of the 16th International Conference on the Kallikrein-Kinin System (Kinin 2002) which was held in Charleston, SC, May 26 – 31, 2002 (see International Immunopharmacology Volume 2/13 – 14). * Corresponding author. Tel.: +1-631-444-2352; fax: +1-631-444-2493. E-mail address: [email protected] (B. Ghebrehiwet).
1567-5769/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. doi:10.1016/S1567-5769(02)00270-9
300
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
1. Introduction Surface properties of endothelial cells play a critical role in maintaining vessel wall integrity and in the pathophysiology of thrombosis, atherosclerosis and inflammation. This response is largely in part due to their ability to respond to a wide range of environmental signals with specific programmed responses (reviewed in Ref. [1]). Certain endothelial cell responses are rapid and do not require new gene transcription or protein synthesis. These include secretion of Weibel-Palade bodies with concomitant expression of surface P-selectin, activation of nitric oxide (NO) metabolic pathways and cytoskeletal or junctional rearrangement resulting in increased vascular permeability [1]. These responses are the hallmark of the early non-leukocyte-dependent phase of inflammation and a large number of agonists are capable of eliciting this type of response including proteins of the clotting and complement systems [1]. Various receptor systems have also been described on vascular cells that participate in the recognition, activation and clearance of components involved in humoral defense. Among these molecules are receptors for the collagen-like tail [cC1q-R or collectin receptor, a homologue of calreticulin (CR)] [2 – 4], and globular heads, gC1q-R/p33 (also known as p32) [5,6], of the complement component C1q. Interaction of C1q with endothelial cells in turn is known to induce a diversity of biological functions including adhesion and spreading [5]; stimulation and expression of the adhesion molecules E-selectin, ICAM-1 and VCAM-1 [7]; and production of IL-6, IL-8 and monocyte chemoattractant protein-1 [8]. Since C1q is present in significant quantities at sites of atherosclerotic and inflammatory and vascular lesions, and gC1q-R as well as cC1q-R are present on endothelial cells, modulation of endothelial cell function by soluble and immobilized C1q may contribute significantly to the development of thrombosis and inflammation. In addition, endothelial cell gC1q-R [9,10], together with the urokinase plasminogen activator receptor (u-PAR) [11], and cytokeratin 1 [12], has been shown to serve as a zinc-dependent, high-affinity site for high-molecular weight kininogen (HK) and factor XII (FXII) [9] that leads to the generation of the potent multifunctional peptide, bradykinin [13]. The present studies were therefore undertaken to address
the following specific questions: (a) which C1q receptors and/or other molecules are involved in the C1q-mediated activation of endothelial cells? (b) Does zinc expose neoepitopes on endothelial cell surface so that HK-binding to gC1q-R is enhanced?
2. Materials and methods 2.1. Chemicals and reagents Unless specified elsewhere, the following chemicals and reagents were purchased from the commercial sources indicated: FCS (Hyclone Laboratories, Logan, UT): RPMI 1640, 100 antibiotic antimycotic mixture, Dulbecco’s phosphate buffered saline (D-PBS), GRRGDSP and GRRGESP peptides, as well as anti-a5 integrin (P1D6) (GIBCO BRL, Gaithersburg, MD); mAb (4B4) antibody to h1 integrin (Coulter Immunology, Hialeah, FL). 2.2. Proteins Highly purified human C1q, free of the C1q inhibitor chondroitin 4-sulfate proteoglycan [14], was either purified as described [15] or purchased from Advanced Research Technologies (San Diego, CA). Before use in culture with endothelial cells, the C1q was dialyzed against sterile RPMI or D-PBS before use in culture with endothelial cells. Heatinactivated C1q was prepared by incubating C1q (1 mg/ml) at 56 jC for 1 h. The residual activity of inactivated C1q (DC1q) was monitored by comparing its activity with untreated C1q by a standard hemolytic assay using C1q-depleted (C1qD) serum as described earlier [16]. 2.3. Antibodies The production and characterization of monoclonal antibodies (mAb) 60.11 and 74.5.2 as well as polyclonal antibodies (pAb) to gC1q-R and to cC1q-R in rabbits have been described earlier [17]. The mAb 60.11 is directed against the N-terminal segment containing a C1q-binding site, whereas the 74.5.2 recognizes epitopes in a segment of gC1q-R containing a site for high-molecular weight kininogen. The IgG fraction was purified as described in detail
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
previously [18]. Monoclonal antibody specific for h1 integrin (mAb 4B4) was purchased from Coulter Immunology and anti-a5 integrin (mAb P1D6) was purchased from GIBCO BRL. The following monoclonal antibodies: A3D8 (anti-CD 44, Mr 80 – 110 kDa), AF3 (anti-CD44H, Mr 100 kDa), P1B5 (antia3 integrin), a4 (anti-a4 subunit of VLA 4) and P1D4 (anti-a5 subunit of integrin) were a generous gift from Dr. Michael Shepley, Division of Infectious Diseases, SUNY at Stony Brook. 2.4. Culture of human dermal microvascular endothelial cells (HDMVECs) Human dermal microvascular endothelial cells were isolated from human neonatal foreskins as previously reported [19]. Briefly, after initial harvest from minced trypsinized human foreskins, microvascular endothelial cells were further purified on a Percoll density gradient. HDMVECs were cultured on collagen type 1 coated tissue culture flasks in endothelial cell growth medium (EGM) consisting of endothelial basal medium (EBM) supplemented with 10 ng/ml epidermal growth factor, 0.4% bovine brain extract, 17.5 Ag/ml dibutyryl cyclic AMP and 1 Ag/ml hydrocortisone in the presence of 30% normal human serum. Endothelial cell cultures were characterized and determined to be >99% pure on the basis of formation of typical cobblestone monolayers in culture, positive immunostaining for von Willebrand factor and selective uptake of acetylated low-density lipoprotein. All experiments were done with HDMVECs below passage 10 as described in our recent paper [20]. 2.5. Endothelial cell adhesion and spreading assay The assay used here for measuring endothelial cell adhesion and spreading was modified from a previously described standard assay [21]. Immulon-4 microtiter plates (Dynatech Laboratories, Chantilly, VA) were coated with 50 Al/well of C1q (10 –50 Ag/ ml) or other purified proteins diluted in 50 mM Tris pH 7.4, 150 mM NaCl and 0.1% NaN3. Bovine serum albumin (BSA, 20 mg/ml) or DC1q were used as negative controls, while type I collagen (10 Ag/ml) was used as positive control. Plates were incubated overnight at 4 jC. All plates were then washed,
301
incubated with 2% BSA in PBS containing 0.1% NaN3 for 2 h at room temperature (20 jC) to block nonspecific binding sites, and then washed again prior to use. HDMVECs were washed twice, harvested, resuspended in assay buffer (EBM with 0.1% BSA) and incubated with or without inhibitors for 30 min at room temperature prior to initiation of the assay. Cell attachment and spreading was measured by adding 100 Al aliquots of HDMVEC suspension (104 cells/ well) to the coated microtiter plate wells and incubating the plate at 37 jC for 1 – 2 h for attachment and 6– 8 h for spreading. To assess the potential effect of antibodies on cell adhesion or spreading, the cells were pre-treated (30 min, 37 jC) with a predetermined concentration of the anti-C1q-R or anti-integrin antibodies before addition to the C1q-coated plates. At the end of the assay, 100 Al of 2% glutaraldehyde was carefully added to each well to fix the attached cells. Then, the unattached cells were removed by gently washing the wells two times with PBS and one time with distilled water. Endothelial cell adhesion and spreading was observed and recorded with a Nikon Diaphot-TMD inverted microscope (Nikon, Melville, NY), equipped with a video system consisting of a Dage-MTI CCD-72S video camera and linked to a Macintosh G3 computer. The images were captured at various magnifications using Adobe Photoshop. All experiments were repeated at least three times. 2.6. Quantification of cell adherence and spreading The percent of adherent cells was quantified spectrophotometrically by the detergent-compatible bicinchoninic acid method after solubilizing the cells with 4,4-dicarboxy-2,2V-biquinoline (Pierce). Nonspecific adhesion to BSA- or DC1q-coated wells was subtracted from each experiment. Inhibition of spreading was calculated by subtracting the number of spreading cells counted under the microscope from the total adherent cells in each experiment as described earlier [20]. 2.7. Cell surface biotinylation of HDMVECs The confluent HDMVECs monolayers (106/ml) grown in a collagen-coated flask were surface biotinylated in situ using the surface impermeable reagent
302
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
sulfo-NHS-LC-biotin as described previously [18,20]. The adherent cells in the flask were washed three times in warm HEPES-buffered saline (HBS) (10 mM HEPES, 137 mM NaCl, 4 mM KCl, 11 mM glucose) and incubated (4 jC, 2 h) with 10 ml of 5 mM sulfoNHS-LC-biotin. After incubation, the excess biotin was removed; the cells were washed twice in 20 ml HBS and lysed using lysis buffer [10 mM HEPES, 150 mM NaCl, 2 mM phenylmethylsulfonyl fluoride (PMSF), 1 AM aprotinin, 1 AM pepstatin, 1 mM EDTA and 0.1% soybean trypsin inhibitor (SBTI) and 1% NP-40] on ice. The cell lysate was finally taken out from the flask, transferred onto sterile test tubes and, after removal of the nuclei and insoluble cellular debris by centrifugation (15 min, 850 g, 4 jC), the supernatant containing the labeled proteins was subjected to further centrifugation (1 h, 45,000 g, 4 jC). The supernatant was then collected, total protein concentration estimated by the BCA protein assay (Pierce) and the degree of biotinylation verified by ELISA using alkaline phosphatase-conjugated streptavidin as a probe. The labeled protein solution was used
either immediately or aliquoted and kept frozen at 80 jC. 2.8. Antigen capture enzyme-linked immunosorbent assay (AC-ELISA) For the AC-ELISA, microtiter plates (MaxiSorb, Nunc, Denmark) were first coated with 100 Al of a 10 Ag/ml (carbonate buffer pH 9.5) of either the capturing polyclonal anti-gC1q-R antibody or non-immune, species-matched control IgG (2 h, 37 jC), washed with TBS-T, blocked with 2% BSA and, after further washing, 100 Al of biotinylated HDMVEC membranes that had been diluted to a concentration of 100 Ag/ml in TB containing 0.5 M NaCl, 0.05% Tween 20 and 0.1% BSA, were added to each well and incubated (overnight, 4 jC). After incubation, the wells were washed once with TB containing 1 M NaCl, 0.05% Tween 20, twice with TBST and once with TBS. Washes in high salt were essential in order to reduce non-specific binding. The captured proteins were then detected using monoclonal antibodies to
Fig. 1. Dose-dependent adhesion of endothelial cells to C1q. Endothelial cells (1 105/well) were incubated (overnight, 4 jC) in 24-well microtiter plate wells coated with either 1 mg/ml BSA, or concentrations of C1q ranging from 10 to 50 Ag/ml. After incubation, the cells were washed three times with PBS, and the adherent cells were fixed and visualized under the microscope. The figure is a representative of four experiments run in duplicates (adapted from Ref. [20]).
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310 Table 1 Effect of anti C1q-R antibodies on endothelial cell spreading and/or adhesion Antibody (30 Ag/ml)
Spreading (% F SDE)
Adhesion (% F SDE)
Anti-gC1q-R (mAb 60.11) Anti-cC1q-R (pAb)
68 F 12 54 F 7
47 F 9 39 F 5
Cells were first pre-treated (30 min, 4 jC) with either 30 Ag/ml of control rabbit IgG, pAb anti-cC1q-R (polyclonal antibody) or mAb 60.11 before addition to C1q-coated wells. The data represent values after subtraction of non-specific adhesion or spreading with IgG controls.
either gC1q-R or integrins and further developed by standard ELISA. Standard for the capture assay included concentrations (0– 1000 ng/ml, in TBS containing 0.5 M NaCl) of highly purified gC1q-R, whereas a similarly treated irrelevant antigen, BSA, was used as control for non-specificity. To ensure that the captured antigens are surface-labeled, a duplicate AC-ELISA was performed under the same conditions except that the captured antigens were detected by alkaline phosphatase-conjugated streptavidin or extravidin and visualized by reaction with p-nitrophenyl phosphate (pNPP).
303
ISS PC1 photon counting spectrophotometer at 25 jC. Each sample contained 0.1 Amol of gC1q-R and 1 Amol of bis-ANS either in HEPES only or with one of the salts at a final concentration of 500 AM: KCl, CaCl2, MnCl2, MgCl2, ZnCl2 and ZnCl2 +EDTA. 2.10. Effect of zinc on endothelial cells To investigate the effect of zinc on surface expressed gC1q-R, flow cytometric analysis was performed using human bone marrow endothelial cells (HBMECs). HBMECs were first pre-incubated with 5 Ag/ml human Fc fragments to block Fc receptors followed by incubation (1 h, 37 jC) with either mAb 74.5.2 alone or in the presence of 50 AM zinc. Isotype- and species-matched non-immune IgG instead of mAb was used as control for non-specificity. After incubation, the cells were washed and the bound antibody probed with Alexa 488-conjugated F(abV)2 goat anti-mouse antibody.
3. Results
2.9. Fluorescence spectroscopy
3.1. Adhesion and spreading of endothelial cells on C1q
Bis-ANS fluorescence experiments were performed with an excitation wavelength at 395 nm and emission wavelength from 430 to 560 nm on
Previous studies have shown that various types of endothelial cells are capable of binding human C1q and that this binding can trigger biological responses
Fig. 2. RGD inhibits C1q-mediated endothelial cell spreading. Cells were incubated (30 min, 4 jC) with either 30 AM RGE or RGD before addition to C1q-coated wells. After incubation, the cells were treated and processed as above. The figure is a representative of four such experiments run in duplicates (adapted from Ref. [20]).
304
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
that include adhesion and spreading. Employing cultured HDMVECs and C1q-coated microplate wells as a model system, we show here that these cells also are able to specifically bind to and spread on C1q (Fig. 1). The binding and spreading was dose-dependent and at physiologic ionic strength, and did not require the presence of metal ions. Furthermore, adhesion to C1q was inhibited when the binding was performed in the presence of (FabV)2 anti-C1q antibody or when ?C1q was used instead of C1q (not shown). No adhesion or spreading occurred on wells that had been coated with 1 mg/ml BSA. The adhesion of HDMVECs to C1q was found to be quantitatively and qualitatively similar to that of collagen type 1 (Fig. 1). 3.2. Spreading of HDMVECs to C1q is inhibited by anti-gC1q-R and anti-cC1q-R To investigate whether the endothelial cell binding is mediated by cC1q-R and/or gC1q-R, HDMVECs were first pre-incubated (30 min, 4 jC) with concentrations ranging from 10 to 100 Ag/ml of either control IgG or mAb 60.11 before addition to C1q-coated wells. After incubation, the cells were washed and the adherent cells fixed and visualized as described. Table 1 shows that spreading (68% F 12) and, to a moderate degree (47% F 9), adhesion was inhibited by mAb 60.11. This inhibition was dose-dependent but no inhibition was noted with even the highest dose (100 Ag/ml) of control IgG. Similar results were obtained when endothelial cells were pre-incubated with 30 Ag/ml pAb anti-cC1q-R antibody (Table 1). When the two antibodies were mixed in the proportion of 100 (60.11) and 30 Ag/ml (pAb anti-cC1q-R), they had a slight (75% F 13) additive effect in their ability to inhibit spreading. 3.3. RGD inhibits C1q-mediated endothelial cell spreading Using human diploid fibroblasts, it has been shown previously that adhesion of cells to C1q was inhibited by soluble GRGDTP peptide [22]. This suggested that adhesion of cells to C1q may require the participation of C1q receptors and integrins. To test this hypothesis, HDMVECs were incubated (30 min, 4 jC) with either 30 AM GRRGDSP (RGD) or GRRGESP (RGE)
Fig. 3. Effect of monoclonal antibodies to h1 and a5 integrins on endothelial cell spreading. Endothelial cells were first pretreated (30 min, 4 jC) with either 50 Ag/ml control IgG or mAb anti-b1 integrin. The cells (1 105/well) were then added to wells coated with C1q, and visually examined for spreading and adhesion as described above.
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
305
before addition to C1q-coated wells. After incubation, cells were treated and processed as described above. As shown in Fig. 2, RGD peptide but not RGE was able to completely (100%) inhibit endothelial cell spreading but not adhesion as assessed by visual examination and manual counting. Those that remained adherent were round and without any of the characteristic cytoskeletal reorganization and formation of typical cobblestone monolayers seen on spreading cells.
sion as described above. Fig. 3 is a representative experiment and shows that, while mAb anti-h1 integrin inhibited both spreading and adhesion of HDMVECs to C1q-coated wells, anti-a5 integrin (not shown) was able to inhibit only cell spreading. Additional experiments, which were similarly designed to test the effect of other integrins, had minimal or no significant inhibitory effect.
3.4. Effect of monoclonal antibody to b1 integrin on endothelial cell spreading
The fluorescence intensity of bis-ANS greatly increases upon binding to hydrophobic sites and has been used widely to assess the hydrophobic surface of proteins [23]. A blue shift in emission maximum is also observed upon binding of bis-ANS to hydrophobic surface in comparison to emission maximum of 520 nm for free bis-ANS. Fig. 4 shows cationinduced change of bis-ANS fluorescence by gC1q-R changing the bis-ANS emission spectra from 520 to 485 nm. Addition of zinc further blue-shifted the bisANS emission spectra from 485 to 475 nm and produced three- to four-fold enhancement in fluorescence intensity. Addition of EDTA to the solution containing gC1q-R, bis-ANS and zinc results in a four-fold decrease in the bis-ANS fluorescence intensity indicating that the zinc-induced hydrophobic exposure is reversible. Furthermore, it was demon-
To identify the type of integrin(s) that may be involved in C1q-mediated endothelial cell spreading, AC-ELISA was first performed on HDMVEC membrane proteins using pAb anti-gC1q-R-coated wells. The captured proteins were then probed with various anti-integrin antibodies as described in Section 2. The results of this experiment showed that h1 integrin is cocaptured (not shown) with gC1q-R giving strong evidence that h1 integrins and gC1q-R may collaborate to induce C1q-mediated spreading. To test this hypothesis, HDMVECs were first pre-treated (30 min, 4 jC) with either 50 Ag/ml control IgG or mAb anti-h1. The cells (1 105/well) were then added to C1q-coated wells and visually examined for spreading and adhe-
3.5. Zinc can induce hydrophobic exposure in gC1q-R
Fig. 4. Bis-ANS fluorescence emission spectra. Emission spectra from bottom to top are: (i) ZnCl2 + EDTA, CaCl2, buffer alone, ZnCl2.
306
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
Fig. 5. Flow cytometric analysis of endothelial cells. Endothelial cells were reacted with mAb 74.5.2 in the presence [3] or absence [2] of 50 AM ZnCl2. The bound antibody was probed with Alexa-488conjugated F(abV)2 goat anti-mouse antibody. Species- and isotypematched IgG (MOPC) was used as control for non-specificity [1]. The figure is a representative of five experiments.
strated that the zinc-induced exposure of hydrophobic surface is not the result of increased nonspecific ionic strength, since the fluorescence intensity of bisANS + gC1q-R did not increase in the presence of other divalent or monovalent cations, such as MgCl2, MnCl2, CaCl2 and KCl. 3.6. Endothelial cells express enhanced antibody binding in the presence of zinc Endothelial cells express gC1q-R on their surface as evidenced by moderate binding of either mAb 60.11 or 74.5.2. However, while the binding of mAb 74.5.2 to endothelial cells is enhanced in the presence of zinc (Fig. 5), no increase in binding was noted with mAb 60.11 (not shown), which recognizes an epitope in the N-terminal region of gC1q-R under the same experimental conditions. No significant staining was also observed with either control IgG alone or in the presence of zinc (not shown). Similar results were also observed with platelets and the cell line U937.
4. Discussion The ability of C1q to interact with various types of vascular endothelial cells leading to the production of biologically active proteins or expression of adhesive
molecules has been well documented. Previously, we and others have shown that endothelial cells constitutively express both cC1q-R/CR and gC1q-R [3– 5] and that the expression of these molecules is upregulated by inflammatory cytokines [6]. However, despite the availability of data implicating these molecules to play a significant role in various C1q-mediated endothelial cell function, the lack of a ‘‘consensus motif’’ in their respective sequences consistent with a transmembrane domain has made it difficult to explain how these molecules communicate with elements inside the cell. The present studies were, therefore, undertaken to explore the possibility that such molecules may circumvent their lack of direct access to the interior of the cell by forging an association with membrane-spanning cell surface proteins in a manner similar to uPAR [24 – 26]. To address this question, we employed an experimental system in which the adherence and spreading of human dermal microvascular endothelial cells on C1q-coated microtiter wells was investigated, and the involvement of the two C1q receptors in question, cC1q-R/CR and or gC1q-R, was assessed. Our results show that HDMVECs, like other endothelial cell types, can adhere and spread on C1q-coated microtiter wells in a specific and dose-dependent fashion (Fig. 1), and this finding is consistent with the long-held postulate that C1q can act as a matrix protein for endothelial cells and fibroblasts [4,5,22]. The degree of adherence and spreading was qualitatively and quantitatively comparable to the adherence of HDMVECs on type I collagen. Furthermore, while adherence was not drastically affected, spreading of HDMVEC to C1q was inhibited when the cells were first pre-treated with either mAb 60.11, an antibody directed to a C1q-binding site on the N-terminus of gC1q-R or polyclonal anti-cC1q-R antibodies, while species- and isotype-matched control IgG had no effect on either adherence or spreading. More importantly, however, spreading was completely and adherence moderately inhibited when HDMVECs were preincubated with 30 AM of soluble RGD peptide but not RGE before addition to C1q-coated wells (Fig. 2). Moreover, pre-treatment of HDMVECs with 50 Ag/ml of mAb anti-h1 integrin before addition to C1q-coated wells also resulted in the inhibition of adhesion and spreading, but pre-incubation of cells with control IgG did not. Of the anti-a integrins tested, only anti-a5
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
integrin was able to moderately inhibit spreading indicating that it may be a likely partner of the h1 integrin. AC-ELISA using anti-gC1q-R also confirmed that h1 integrin is co-captured from solubilized HDMVEC membranes. The studies were undertaken to answer a simple but intriguing biological question: how do proteins like cC1q-R and gC1q R, which lack a traditional membrane spanning domain, communicate with elements inside the cell? The results presented in this report collectively suggest that, at least on HDMVECs, C1q-mediated adhesion and spreading may require the participation of cell surface C1q receptors and h1 integrins, and possibly other molecules. This model would envisage that both cC1q-R and gC1q-R, which form a high-affinity binding complex upon ligand binding [18], would laterally associate with h1 integrin to form a docking/signaling complex. This kind of molecular association between membrane-spanning and non-membrane-spanning proteins to form a docking/signaling complex is not unique to gC1q-R and/or cC1q-R. Many surfaceassociated proteins, which lack transmembrane domains, are known to trigger biological responses using this strategy. For example, urokinase plasminogen activator receptor (CD87), which is also expressed on vascular endothelial cells, is a GPIanchored protein and, as such, has no direct link with signaling proteins inside the cell [24]. However, uPAR can form a complex with h1 or h2 integrins to modulate their adhesive functions and with the h2 integrin, CR3 (CD11b/CD18), to trigger uPA-induced Ca2 + fluxes in neutrophils [24,25]. More recently, a signaling partnership between uPAR and L-selectin (CD62L) was also demonstrated in human polymorphonuclear neutrophils [26]. These findings, therefore, would tend to support the concept that membrane-associated proteins can pick and choose their signaling partners depending on the cell type and signal required. Recent evidence also suggests that C1q and mannose-binding lectin (MBL) can engage cell surface calreticulin (cC1q-R) and CD91 to initiate macropinocytosis and uptake of apoptotic cells. MBL is a member of the family of proteins collectively known as collectins (collagen containing lectins) and, like C1q, binds to cC1q-R or calreticulin (or collectin receptor) [27,28]. Apoptotic cell uptake via the calreticulin/CD91 docking/signaling complex is, therefore,
307
another example of the molecular partnership that is forged between transmembrane and non-transmembrane proteins in the induction of specific biological responses. In addition to its participation in C1q-mediated responses, gC1q-R has been shown to serve as a high-affinity receptor for high-molecular weight kininogen. The binding of HK to gC1q-R is entirely dependent on the presence of zinc [9,10] and is inhibited by mAb 74.5.2, which recognizes an epitope within the gC1q-R segment corresponding to residues 204– 218. This segment, which forms part of the h6 strand in the crystal structure [29], is largely restricted to the inside of the central hole (Fig. 6). The homotrimeric, doughnut-like gC1q-R molecule has two arbitrarily assigned faces [30]. This assignment is primarily based on the large difference in charge between the two faces. One side is highly acidic and is likely going to interact with ligands in solution (Sface), whereas the mildly basic face interacts with the cell membrane (M-face). Although mAb 74.5.2 can recognize gC1q-R on the quiescent endothelial cell surface, its interaction is greatly enhanced in the presence of 50 AM zinc. This is largely because zinc can induce exposure of hydrophobic sites in the Cterminal domain of gC1q-R, which includes residues 204– 218 (Figs. 3 and 4). It is under these conditions we postulate that the binding of HK to the endothelial cell gC1q-R becomes greatly increased and this highaffinity (9 F 2 AM) interaction would then trigger the activation of the kinin-forming cascade. Bradykinin, thus generated, would in turn bind to the endothelial cell bradykinin receptor (B2) and induce a plethora of pathophysiologic responses ranging from tumor metastasis and angiogenesis to the induction of morphologic changes in the endothelium rendering the subendothelial matrix accessible to blood components, thus leading to infiltration by inflammatory cells. Although both cC1q-R and gC1q-R are constitutively expressed on resting nonthrombotic endothelium and gC1q-R in particular has the potential to activate the bradykinin-generating system [13], the mechanism by which a continuously thrombogenic state of the endothelium is averted is not known. Based on the above findings, we speculate that efficient engagement of gC1 q-R by its proinflammatory ligands or C1q-containing immune complexes is re-
308
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
Fig. 6. The localization of antibody binding sites. The localization of mAb 60.11, which recognizes epitopes in the N-terminus containing a C1q-binding site is predominantly exposed on the S face of the molecule, whereas mAb 74.5.2 recognizing the HK site is restricted to the inside of the hole of the homotrimeric molecule.
stricted to conditions where, under chemical, physical or infectious insult, the endothelial cell is converted to a prothrombotic and proinflammatory phenotype leading to the induction of cytokines, such as IL-1 or TNFa or expression of cell adhesion molecules (CAMs) [1]. Leukocytes bound to the cell adhesion molecules can release cytokines, which in turn can amplify the system by upregulating the expression of C1q-binding molecules in a manner that allows efficient binding of ligands such as C1q or HK. That cytokines, such as LPS, IFNg and TNFa, can upregulate the expression of both cC1q-R and gC1q-R has been reported previously [6]. Thus, modulation of
endothelial cell function by soluble and/or immobilized proinflammatory ligands such as C1q or HK especially at sites of inflammatory and vascular lesions may contribute significantly to the development of thrombosis and exacerbation of the inflammatory process.
Acknowledgements This work was supported in part by grants RPG95068-06 from the American Cancer Society (B.G.), R01 HL5029101 from National Heart Blood and
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
Lung Institute (E.I.B.P. and B.G.) and a generous gift from Larry and Sheila Dalzell (B.G.). We are indebted to the late Dr. Michael Shepley, Division of Infectious Diseases, SUNY Stony Brook, without whose generous gift of the anti-integrin antibodies, this work would not have been completed. The expert technical assistance of Dana Jaggerssarsingh and Lynda Piboon is also greatly acknowledged.
References [1] Silverstein RL. The vascular endothelium. In: Gallin JI, Snyderman R, editors. Inflammation: Basic Principles and Clinical Correlates. Philadelphia: Lippincott Williams and Wilkins; 1999. p. 207. [2] Andrews BS, Shadforth M, Cunningham P, Davis IV JS. Demonstration of a C1q receptor on the surface of human endothelial cells. J Immunol 1981;127:1075. [3] Peerschke EIB, Malhotra R, Ghebrehiwet B, Reid KBM, Willis AC, Sim RB. Isolation of a human endothelial cell C1q receptor (C1q-R). J Leukoc Biol 1993;53:179. [4] Daha MR, Dunn L, Van den Berg RH, Muizert-de Lange Y, Gerritsen AF, Van Eys LA. Isolation of a human endothelial cell C1q receptor. Mediat Inflamm 1993;2:447. [5] Peerschke EIB, Smyth SS, Teng E, Dalzell M, Ghebrehiwet B. Human umbilical vein endothelial cells possess binding sites for the globular domains of C1q. J Immunol 1996;157:4154. [6] Guo W-X, Ghebrehiwet B, Weksler B, Schweizer K, Peerschke EIB. Upregulation of endothelial cell binding proteins/ receptors for complement C1q by inflammatory cytokines. J Lab Clin Med 1999;133:541. [7] Lozada C, Levin RI, Huie M, Hirschhorn R, Naime D, Whitlow M, et al. Identification of C1q as a heat-labile serum cofactor required for immune complexes to stimulate endothelial expression of the adhesion molecules E-selectin and intercellular and vascular cell adhesion molecules 1. Proc Natl Acad Sci U S A 1995;92:8378. [8] van den Berg RH, Faber-Krol MC, Sim RB, Daha MR. The first subcomponent of complement, C1q, triggers the production of IL-8, IL-6 and monocyte chemoattractant peptide-1 by human umbilical vein endothelial cells. J Immunol 1998; 161:6924. [9] Joseph K, Ghebrehiwet B, Peerschke EIB, Reid KBM, Kaplan AP. Identification of the zinc-dependent endothelial cell binding protein for high molecular weight kininogen and factor XII: identity with the receptor which binds to the globular ‘heads’ of C1q (gC1q-R). Proc Natl Acad Sci U S A 1996; 93:8552. [10] Herwald H, Dedio J, Kellner R, Loos M, Muller-Esterl W. Isolation and characterization of the kininogen-binding protein p33 from endothelial cells. Identity with the gC1q receptor. J Biol Chem 1996;271:13040.
309
[11] Coleman RW, Pixley RA, Najamunnisa S, Yan W, Wang J, Mazer A, et al. Binding of high molecular kininogen to human endothelial cells is mediated via a site within domains 2 and 3 of the urokinase receptor. J Clin Invest 1997;100:1481. [12] Hasan AAK, Zisman T, Schmaier AH. Identification of cytokeratin 1 as a binding protein and presentation receptor for kininogens on endothelial cells. Proc Natl Acad Sci U S A 1998;95:3615. [13] Joseph K, Ghebrehiwet B, Kaplan AP. Activation of the kininforming cascade on the surface of endothelial cells. Biol Chem 2001;382:71. [14] Kolb WP, Kolb LM, Podack ER. C1q: isolation from human serum in high yield by affinity chromatography and development of a highly sensitive hemolytic assay. J Immunol 1979; 122:2103. [15] Silvestri L, Baker JR, Roden L, Stroud RM. The C1q inhibitor in serum is a chondroitin 4-sulfate proteoglycan. J Biol Chem 1981;256:7383 – 7. [16] Conradie JD, Volanakis JE, Stroud RM. Evidence for a serum inhibitor of C1q. Immunochemistry 1975;12:967 – 71. [17] Ghebrehiwet B, Lu PD, Zhang W, Lim B-L, Eggleton P, Leigh LEA, et al. Identification of functional domains on gC1q-R, a cell surface protein which binds to the globular ‘‘heads’’ of C1q, using monoclonal antibodies and synthetic peptides. Hybridoma 1996;15:333. [18] Ghebrehiwet B, Lu PD, Zhang W, Keilbaugh SA, Leigh LEA, Eggleton P, et al. Evidence that the two C1q-binding membrane proteins gC1q-R and cC1q R associate to form a metal ion-independent complex. J Immunol 1997;159:1429. [19] Feng X, Clark RAF, Galanakis D, Tonnesen MG. Fibrin and collagen differentially regulate human dermal microvascular endothelial cell integrins: stabilization of av/b3 mRNA by fibrin. J Invest Dermatol 1999;113:913. [20] Feng X, Tonnesen MG, Peerschke EIB, Ghebrehiwet B. Cooperation of C1q receptors and integrins in C1q-mediated endothelial cell adhesion and spreading. J Immunol 2002; 168:2441 – 8. [21] Gailit J, Clark RAF. Studies in vitro on the role of av and h1 integrins in the adhesion of human dermal fibroblasts to provisional matrix proteins fibronectin, vitronectin, and fibrinogen. J Invest Dermatol 1996;106:102. [22] Bordin S, Ghebrehiwet B, Page RC. Participation of C1q and its receptor in adherence of human diploid fibroblast. J Immunol 1990;145:2520. [23] Cardamone M, Puri NK. Spectrofluorimetric assessment of the surface hydrophobicity of proteins. Biochem J 1992;282: 589 – 93. [24] Sitrin RG, Pan PM, Harper HA, Blackwood RA, Todd III RF. Urokinase receptor (CD87) aggregation triggers phosphoinositide and intracellular calcium mobilization in mononuclear phagocytes. J Immunol 1999;163:6193. [25] Sitrin RG, Todd III RF, Petty HR, Brock TG, Shollenberger SB, Abrecht E, et al. The urokinase receptor (CD87) facilitates CD11b/CD18-mediated adhesion of human monocytes. J Clin Invest 1996;97:1942. [26] Sitrin RG, Pan PM, Blackwood RA, Huang J, Petty HR. Evidence for signaling partnership between urokinase recep-
310
B. Ghebrehiwet et al. / International Immunopharmacology 3 (2003) 299–310
tors (CD87) and L-selectin (CD62L) in human polymorphonuclear neutrophils. J Immunol 2001;166:4822. [27] Sim RB, Moestrup SK, Stuart GR, Lynch NJ, Lu J, Schwaeble WJ, et al. Interaction of C1q and the collectins with the potential receptors calreticulin (cC1qR/collectin receptor) and megalin. Immunobiology 1998;199:208. [28] Malhotra R, Lu J, Thiel S, Jensenius J-C, Willis AC, Sim RB. C1q receptor (collectin receptor): primary structure, ho-
mology, and interaction with ligands. Immunobiology 1992; 184:1437. [29] Jiang J, Zhang Y, Krainer AR, Xu RM. Crystal structure of human p32, a doughnut-shaped acidic mitochondrial matrix protein. Proc Natl Acad Sci U S A 1999;96:3572 – 7. [30] Ghebrehiwet B, Jesty J, Peerschke EIB. gC1q-R/p33: structure – function predictions from the crystal structure. Immunobiology 2002;205:421.