Complement-induced procoagulant platelets in patients with antiphospholipid antibodies

Abstract

Abstract Introduction Antiphospholipid syndrome (APS), an autoimmune disease characterized by persistent circulating antiphospholipid antibodies (aPL), is the most common cause of acquired thrombophilia. Being the pivotal cells in mediating thrombosis, platelets are known to be activated or their responses to sub-threshold agonist potentiated by aPL as well as anti-β2-glycoprotein-1 IgG. We have earlier demonstrated complement activation in thrombotic APS and shown that patients with APS, particularly catastrophic APS, have mutations in complement regulatory genes. Elevated levels of C3a and C4a have been described in APS, but their association with thrombosis is inconsistent. We also previously reported that the C5 inhibitor eculizumab prevented thrombosis in a patient with refractory APS, and other reports demonstrate its efficacy as salvage therapy for CAPS. APS is thus characterized by both increased activation of platelets and the complement cascade. We explore here the contributions of complement-platelet axis to the risk of thrombosis in primary APS. Methods Venous blood was drawn from 40 patients with antiphospholipid antibodies (aPL+) (29/40 had thrombosis - APS, 13/40 were triple positive and 12/40 double positive for aPL, 18/40 were on antiplatelet drug) and 15 healthy controls. Washed platelets were prepared by differential centrifugation. Platelets unstimulated or treated with thrombin, collagen or ADP for 5 min were labelled with FITC-PAC1, PE-anti-CD62P, Alexa Fluor488-anti-C4d, PE-Annexin V, Mitotracker Red, Rhod2-AM, Caspase 3/7 detection reagent or FITC-IETD-FMK for 30 min in the dark, and then analyzed by flow cytometry. Platelet markers were correlated with complement binding and clinical characteristics of APS patients including number of thrombotic events and circulating levels of aPL. In another set of experiments (n=3-6), washed platelets from healthy donors were supplemented with 2.5 mM calcium and were incubated with sera (10%) from APS patients or healthy donors for 1 or 2h in the presence of PE-Annexin V and APC-anti-CD62P along with either prothrombin or anti-C5b-9/C4d/C1q antibodies prelabelled with Alexa-Fluor488 for simultaneous analysis of procoagulant platelets, complement or prothrombin binding. Platelets were then fixed and analyzed by flow cytometry. On certain occasions sera were heat inactivated at 56º C for 30 min or pre-treated with IdeS (IgG cleaving enzyme) (10000 U/ml at 37º C for 30 min), eculizumab (1 mg/ml), sutimlimab (C1s inhibitor) (1.5 mg/ml) and/or danicopan (Factor D inhibitor) (100 µM). Also, on other occasions platelets were pre-incubated with antagonists of C3aR (SB290157), FcgRIIa (IV.3) or C5aR1 (Avacopan) (10 µM) before serum. Results We found that although activation and responsiveness to agonists of platelets from aPL+/APS patients were like those from healthy individuals, unstimulated platelets from APS patients exhibited higher phosphatidylserine (PS) exposure (19.6±3.9% vs 4.0±1.0% controls) and P-selectin expression (15.9±2.3% vs 8.6±1.8% controls) than controls suggesting induction of procoagulant activity. While platelet PS exposure in aPL+/APS was associated with increased mitochondrial calcium, this was not accompanied by mitochondrial depolarization or caspase activation. Platelet PS exposure in APS patients is strongly associated with platelet surface complement C4d deposition. Incubating healthy donor platelets with sera but not aPL obtained from APS patients reproduces the procoagulant platelet phenotype characteristic of APS. APS serum induced procoagulant activity was associated with increased prothrombin and complement binding to platelets. Platelet procoagulant activity as well as complement deposition induced by APS serum was significantly inhibited by prior heat inactivation or pre-incubation with IdeS, sutimlimab, danicopan or eculizumab. Pre-incubating platelets with FcgRIIa, C3aR or C5aR1 antagonists did not reverse the procoagulant activity induced by APS serum. Conclusion We demonstrate a role for complement activation initiated by IgG through the classical complement pathway and amplified by alternative pathway in cooperatively mediating platelet procoagulant activity through C5b-9 deposition on platelets. Finally, we show that procoagulant complement-bound platelets were strongly associated with occurrence of thrombosis in APS patients and could be more effective predictive biomarkers of thrombotic APS than conventional antibody profiling.