Complement mediates binding and procoagulant effects of ultralarge HIT immune complexes

Abstract

Heparin-induced thrombocytopenia (HIT) is a prothrombotic disorder mediated by ultra-large immune complexes (ULICs) containing immunoglobulin G (IgG) antibodies to a multivalent antigen composed of platelet factor 4 and heparin. The limitations of current antithrombotic therapy in HIT supports the need to identify additional pathways that may be targets for therapy. Activation of FcγRIIA by HIT ULICs initiates diverse procoagulant cellular effector functions. HIT ULICs are also known to activate complement, but the contribution of this pathway to the pathogenesis of HIT has not been studied in detail. We observed that HIT ULICs physically interact with C1q in buffer and plasma, activate complement via the classical pathway, promote codeposition of IgG and C3 complement fragments (C3c) on neutrophil and monocyte cell surfaces. Complement activation by ULICs, in turn, facilitates FcγR-independent monocyte tissue factor expression, enhances IgG binding to the cell surface FcγRs, and promotes platelet adhesion to injured endothelium. Inhibition of the proximal, but not terminal, steps in the complement pathway abrogates monocyte tissue factor expression by HIT ULICs. Together, these studies suggest a major role for complement activation in regulating Fc-dependent effector functions of HIT ULICs, identify potential non-anticoagulant targets for therapy, and provide insights into the broader roles of complement in immune complex-mediated thrombotic disorders.

Graphical abstract

Figure 1.

Physical interactions of HIT ULICs and C1q. (A) C1q increases the size of HIT ULICs. PF4 (10 µg/mL) and heparin (H; 0.1 U/mL) were incubated alone or with KKO (50 µg/mL) to generate ULICs. C1q was added to KKO ULICs at molar ratio of 1:4, and size of the complexes was measured by dynamic light scattering. The mean ± standard error of the mean of 3 independent experiments is shown. P < .001 for PF4/heparin + KKO + C1q vs all other conditions starting at 1 hour (2-way analysis of variance). (B) C1q restores formation of ULICs. PF4 (10 µg/mL), KKO (30 µg/mL), RTO (350 µg/mL), and heparin (0.2 U/mL) were preincubated, and buffer alone or buffer containing C1q (1:16 ratio with KKO) was added 1 hour later. The percentage of small particles (<20 nm) is shown at the indicated times. The mean ± standard error of the mean of 3 experiments is shown.

Figure 2.

Complement activation by KKO and HIT ULICs occurs in plasma and WB. Plasma (A,C) or WB (B,D) were incubated with KKO or ISO (A,B) or anti-PF4/heparin IgG from patients with (“HIT IgG”; n = 6) or without (“αPF4/H IgG”, n = 3) HIT or control IgG either in buffer or with PF4/heparin (H) as shown on the x-axis. Complement degradation product C3c was measured. Results shown are representative of 3 independent experiments. ***P < .0001, with all other conditions (A,B) and *P < .05 (C,D) (1-way analysis of variance with Tukey’s multiple comparison test). ns, not significant.

Figure 3.

Effect of inhibitors targeting the proximal steps in the complement pathway in the presence of KKO ULICs. WB was preincubated with the complement inhibitors or respective controls (ISO=isotype or CON=control) shown on the x-axis before adding KKO ULIC, and complement activation was measured by generation of C3c (A) or C5b-9 (B). Results shown are representative of 3 independent experiments. ***P < .0001 (1-way analysis of variance with Tukey’s multiple comparison test). ns, not significant.

Figure 4.

KKO/HIT ULICS activate complement and lead to deposition of complement on neutrophils, monocytes, and B cells. Main figures show binding (mean fluorescence intensity [MFI]) of α-C3c (top) or α-mouse IgG (bottom) to defined cell populations after incubating WB with conditions indicated on the x-axis and analyzed by flow cytometry. Cell populations were gated by labeled Abs to CD66 (neutrophils), CD14 (monocytes), and CD19 (B cells). Insets for each graph show respective cell population when WB was incubated with buffer (B) or HIT IgG (HIT) alone or with PF4/heparin (P+H+HIT) or ISO/control with PF4/heparin (P+H+ISO). The graphic is representative of ≥3 independent experiments using KKO and complement-activating HIT IgG. **P < .005, ***P < .0001 (compared with all other conditions using 1-way analysis of variance with Tukey’s multiple comparison test).

Figure 5.

HIT ULICs promote monocyte procoagulant activity and TF expression in a complement-dependent manner. (A) C1q binding to HIT ULICs enhances monocyte procoagulant activity. KKO/ISO ULICs and human C1q were added to monocytes in serum-free media for 5 hours at 37°C under 5% carbon dioxide. C1q was added at C1q:KKO molar ratios of 1:8 [C1q (hi)] and 1:800 [C1q (lo)]. Aliquots from the cell suspensions were analyzed for FXa activity by using a chromogenic assay. Results of 2 independent experiments, each done in quadruplicate (mean ± standard error of the mean), are shown as a fold increase in FXa generation relative to cells incubated with media alone. ***P < .0001 (compared with all other conditions using the 1-way analysis of variance with Tukey’s multiple comparison test). (B) HIT ULICs induce monocyte TF expression upstream of C5: WB blood was preincubated without or with complement inhibitors/controls (α-C1q ISO Ab/α-C1q Ab 300 μg/mL, CP40-CON/Cp40 20 μM, IV.3-ISO/IV.3 10 μg/mL, α-C5-ISO ab/α-C5 Ab 50 μg/mL) before adding PF4/heparin/KKO ULICs as shown on the x-axis. After 4 hours of incubation, RBCs were lysed, leukocytes were stained with α-CD14 and α-TF Ab, and the percentage of TF-expressing monocytes was measured. The graphic shows the results of 1 of 3 independent experiments tested in duplicate (mean ± SD). ***P < .0001 (1-way analysis of variance with the Tukey’s multiple comparison test). ns, not significant.

Figure 6.

Opsonization by complement facilitates ULIC binding and thereby FcR effector functions. (A) Complement inhibition by α-C1qAb reduces deposition of KKO ULICs. WB was preincubated with buffer, α-C1q (300 µg/mL), or ISO control (300 µg/mL) before adding ULICs containing KKO or ISO IgG (50-200 µg/mL). Surface-bound mouse (m)IgG was measured by flow cytometry. α-C1q-Ab, but not its ISO control, prevented deposition of mIgG on the cell surface. Effects of complement inhibition could not be overcome with higher doses of ULIC-IgG. (B) Complement inhibition inhibits TF expression: Similar incubation conditions from panel A were used, followed by detection of monocyte TF by flow cytometry. Inhibition of complement reduced TF expression. The graphic shows results representative of 3 independent experiments tested in duplicate (mean ± SD). P < .0001 (1-way analysis of variance with the Tukey’s multiple comparison test for KKO+PF4/heparin [H] at each concentration vs each concentration of ISO controls and similarly for conditions using α-C1q at each concentration vs each concentration of ISO controls for α-C1q). *Buffer. **PF4/KKO (200 µg/mL). MFI, mean fluorescence intensity.

Figure 7.

Inhibiting the proximal steps in the complement pathway impairs platelet adhesion to human umbilical vein endothelial cells induced by KKO. Microfluidic channels coated with human umbilical vein endothelial cells were perfused with recalcified citrated WB stimulated with monoclonal Ab KKO (50 µg/mL) in the presence of PF4 (25 µg/mL). Some samples were preincubated with C1r inhibitor [C1r(i); 5 µg/mL, 25 µg/mL, or 100 µg/mL] 15 minutes before the addition of KKO (50 µg/mL) and PF4 (25 µg/mL). Adhesion of calcein-labeled platelets was measured as total fluorescent intensity at 5-minute intervals, and the effect of C1r(i) is expressed as percentage of platelet adhesion in the presence of inhibitor compared with platelet adhesion without inhibitor (100%). *P < .05, **P < .005, ***P < .0005 (2-way analysis of variance with Dunnett’s correction); n = 4.