Antiphospholipid syndrome (APS) is the newest member of the PF4 immunothrombotic family: Mechanistic and therapeutic insights

Abstract

Abstract Background: APS is an autoimmune thromboinflammatory disorder characterized by venous, arterial, and microvascular thrombosis, adverse pregnancy outcomes and thrombocytopenia. A major antigenic target is b2-glycoprotein I (b2GPI). Prior observations showed: i. b2GPI can bind the platelet chemokine, platelet factor 4 (PF4), which enhances binding of APS antibodies (Abs), ii. b2GPI binds extruded chromatin from neutrophils (neutrophil extracellular traps or NETs), which also enhances APS Ab binding, and iii. PF4 binds avidly to NETs. In other immunothrombotic disorders, such as heparin-induced thrombocytopenia (HIT), PF4:NET complexes contribute to thrombosis. Aim: We proposed that PF4 released from platelets binds to NETs, enhancing the binding of b2GPI to the NETs and that PF4:b2GPI:NET complexes are important in the pathobiology of thrombosis in APS. Methods & Results: Dynamic light scattering (DLS) studies showed PF4 and b2GPI form antigenic complexes with DNA that bind isolated APS IgG. These findings were confirmed in a NET-lined microfluidic system, where the NETs bound 5-10 times more b2GPI in the presence of PF4 (25 µg/ml) than in its absence (p<0.0001) and subsequently enhanced APS IgG binding four-fold (p<0.005). IgG from 16 randomly selected individuals who met clinical criteria for APS and had “triple-positive” serology required the addition of β2GPI and PF4 to initiate platelet, fibrin, C1q, and C5b-9 deposition on injured human umbilical vein endothelial cell (HUVEC)-lined microfluidic channels. This effect was not seen with IgG from healthy controls. In vivo studies in mice that express PF4: wildtype (WT) mice, mice expressing only human PF4 (hPF4+) or hPF4 plus FcgRIIA (hPF4+/FcgRIIA+) were compared to mice lacking murine PF4 (mPF4-/-). We found that passive inoculation with APS IgG slowed neutrophil rolling on the uninjured venous endothelium and enhanced neutrophil adhesion and enhanced platelet- and neutrophil-rich thrombosis in laser-injured cremaster venules and arterioles in all mice EXCEPT mPF4-/- mice, which were comparable to mice infused with healthy-donor control IgGs. hPF4+/FcgRIIA+ mice had more neutrophils and neutrophil-derived extracellular vesicles incorporated within venule and arteriole thrombi as compared to mPF4-/- mice. As an independent approach, we tested three anti-PF4 monoclonal (mo) Abs, each having different specificities and effects on PF4: RTO, which disaggregates PF4 tetramers; deglycosylated (DG) 1E12, which blocks PF4 binding to NETs; and an IgG4 version of KKO (G4KKO), which crosslinks PF4:NETs. DLS studies showed that RTO and 1E12 disaggregated PF4:b2GPI:DNA immune complexes, whereas G4KKO formed larger complexes that sterically blocked binding of APS IgG. All 3 Abs inhibited thrombosis in the injured HUVEC system. In mice, G4KKO prevented enhanced neutrophil-venous endothelial interaction when infused prophylactically. G4KKO blocked thrombosis in both venule and arteriole models, when given prophylactically or therapeutically post-APS IgG exposure. Conclusions: PF4:β2GPI:NETs complexes are recognized by triple-positive APS IgG, as demonstrated by all 16 APS IgG samples tested to date. This finding reveals a previously unrecognized similarity between APS and other PF4-mediated immunothrombotic disorders, such as HIT and vaccine-induced immune thrombotic thrombocytopenia (VITT). Notably, each of the three anti-PF4 monoclonal antibodies — each disrupting PF4:β2GPI:NET immune complexes — significantly reduced thrombosis, particularly when administered prophylactically. As these moAbs do not alter hemostasis, they are a potential complementary therapy to standard anticoagulants. Future studies will address the importance of PF4 in other forms of APS that do not involve high-titer anti-β2GPI antibodies.