Cofactor regulation of C5a chemotactic activity in physiological fluids. Requirement for the vitamin D binding protein, thrombospondin-1 and its receptors

Abstract

Factors in physiological fluids that regulate the chemotactic activity of complement activation peptides C5a and C5a des Arg are not well understood. The vitamin D binding protein (DBP) has been shown to significantly enhance chemotaxis to C5a/C5a des Arg. More recently, platelet-derived thrombospondin-1 (TSP-1) has been shown to facilitate the augmentation of C5a-induced chemotaxis by DBP. The objective of this study was to better characterize these chemotactic cofactors and investigate the role that cell surface TSP-1 receptors CD36 and CD47 may play in this process. The chemotactic activity in C-activated normal serum, citrated plasma, DBP-depleted serum or C5 depleted serum was determined for both normal human neutrophils and U937 cell line transfected with the C5a receptor (U937-C5aR). In addition, levels of C5a des Arg, DBP and TSP-1 in these fluids were measured by RIA or ELISA. Results show that there is a clear hierarchy with C5a being the essential primary signal (DBP or TSP-1 will not function in the absence of C5a), DBP the necessary cofactor and TSP-1 a dependent tertiary factor, since it cannot function to enhance chemotaxis to C5a without DBP. Measurement of the C5a-induced intracellular calcium flux confirmed the same hierarchy observed with chemotaxis. Moreover, analysis of bronchoalveolar lavage fluid (BALF) from patients with the adult respiratory distress syndrome (ARDS) demonstrated that C5a-dependent chemotactic activity is significantly decreased after anti-DBP treatment. Finally, results show that TSP-1 utilizes cell surface receptors CD36 and CD47 to augment chemotaxis, but DBP does not bind to TSP-1, CD36 or CD47. The results clearly demonstrate that C5a/C5a des Arg needs both DBP and TSP-1 for maximal chemotactic activity and suggest that the regulation of C5a chemotactic activity in physiological fluids is more complex than previously thought.

Figure 1

Chemotactic activity of C-activated fluids. Pooled human serum, citrated plasma, DBP-depleted serum, C5-depleted serum were treated with zymosan A for 45 min at 37°C to activate complement, and the activated fluid was centrifuged to remove zymosan particles. Panel A: U937-C5aR cells (3×10⁵) were allowed to migrate toward 2.5% dilution of each fluid for 120 min at 37°C. Panel B: Neutrophils (2×10⁵) were allowed to migrate toward 2.5% dilution of each fluid for 25 min at 37°C. Numbers represent mean migration distance ± SEM, n = 3–7. Statistical significance is indicated.

Figure 2

Effect of DBP depletion on the C5a dependent chemotactic activity in ARDS BALF. The indicated ARDS BALF samples (normalized for total protein content), as well as 1 nM purified human C5a, were treated for 30 min at 22°C with 40 µg/ml of goat anti-DBP or an irrelevant goat IgG (sham control). U937-C5aR cells (3×10⁵) were then allowed to migrate towards either C5a or a 20% dilution of BALF in chemotaxis buffer for 120 min at 37°C. Numbers represent mean migration distance ± SEM, n = 3. Asterisks denote that cell movement was significantly less (p < 0.01) than the sham control.

Figure 3

Effect of DBP and TSP-1 on C5a-induced intracellular calcium mobilization in U937-C5aR cells. Panel A: Cells treated with 1% (diluted in HBSS) C-activated or sham activated serum, citrated plasma or DBP-depleted serum. Panel B: Cells were treated with purified C5a, DBP and TSP-1 with no preincubation. Panel C: Cells were preincubated with either DBP or TSP-1 before stimulation with C5a. Cells (5×10⁶ cells/ml) loaded Fluo-3 AM were washed and then suspended in HBSS containing 1% BSA. Cells were pretreated for 20 min at 22°C with either 0.5 nM TSP-1 or 50 nM DBP. For each measurement, 0.4 ml of cells was added to a cuvette and stimulated with either 1% serum or plasma (panel A) or 0.1 nM purified C5a, C5a + 0.5 nM TSP-1 or C5a + 50 nM DBP (panels B and C) at 22°C. Fluorescence was measured immediately at 505 nm excitation, 526 nm emmission for Fluo-3 AM. F_max was measured by treating labeled cells with 60 µM digitonin. Data are presented as mean ± SEM (n = 4–6) of the peak increase in intracellular calcium concentration. Significance is indicated in panel A, double asterisks (panel C) indicates value is significantly higher (p < 0.01) than C5a alone.

Figure 4

Ligand binding to DBP immobilized on a Biacore SPR sensor chip. Panel A: 10⁶ cells/ml in HBSS were not treated (control), pretreated with 50 nM DBP, or with 25 µg/ml of the indicated goat polyclonal antibody. Panel B: Cells were treated with 50 µg/ml of either anti-CD36 or anti-CD47 Fabs. Panel C: binding to immobilized DBP by soluble goat anti-DBP (5 µg/ml), purified TSP-1 (50 nM) or purified C5a (0.5 µM). U937-C5aR cells at 10⁶ cells/ml (panels A and B), or the indicated purified proteins (panel C), were allowed to interact with DBP immobilized on a Biacore sensor chip at a flow rate of 5 µl/min at 22°C in HBSS. Data is expressed as response units of the molecular interaction on the sensor chip.

Figure 5

Effect of anti-CD36 and anti-CD47 on chemotaxis to C-activated serum. Panel A: U937-C5aR cells, Panel B: neutrophils. U937-C5aR cells (3×10⁵) or neutrophils (2×10⁵) in chemotaxis buffer were either not treated (control) or treated for 15 min at 22°C with 25 µg/ml of goat anti-CD36, anti-CD47 or anti-gC1qR. Cells were then allowed to respond to either 2.5% C-activated serum or 1 nM purified C5a for either 25 min (neutrophils) or 120 min (U937 cells) at 37°C. Numbers represent mean migration distance ± SEM, n = 4–6. Double asterisks denote that cell movement was significantly less (p < 0.01) than to the untreated control.

Figure 6

Effect of different anti-CD36 and anti-CD47 on chemotaxis to C-activated serum. U937-C5aR cells (3×10⁵) in chemotaxis buffer were either not treated (control) or treated for 15 min at 22°C with 25 µg/ml of the indicated goat anti-CD36 and anti-CD47 antibodies. Antibody + peptide samples, aliquots of anti-CD36 and CD47 (25 µg/ml) were treated with 40-fold molar excess of antigenic blocking peptide for 15 min at 22°C before they were added to cells. Cells were then allowed to respond to 2.5% C-activated serum for 120 min at 37°C. Numbers represent mean migration distance ± SEM, n = 3. Double asterisks denotes that cell movement was significantly less (p < 0.01) than the control.

Figure 7

Role of CD36 and CD47 in U937-C5aR chemotaxis to C-activated serum versus C-activated plasma. Panel A: Effect of univalent receptor ligation using Fab fragments, Panel B: Effect of peptides to the TSP-1 binding sites. U937-C5aR cells (3×10⁵) in chemotaxis buffer were either not treated (control) or treated for 15 min at 22°C with 25 µg/ml of intact IgG to CD36 or CD47 or 50 µg/ml of the corresponding Fab fragments (panel A) or 100 µg/ml of the CD36 peptides (93–110, 139–155) or the TSP-1 peptide agonist for CD47 (4N1K). Cells were then allowed to respond to either 2.5% C-activated serum or 2.5% C-activated plasma for 120 min at 37°C. Numbers represent mean migration distance ± SEM, n = 4–5. Asterisks denote that cell movement was significantly less (**p < 0.01 or *p < 0.05) compared to the untreated control.