Tissue Factor (TF) Activation by Antithymocyte Globulin Is Inhibited by a Novel Monoclonal Antibody Specific for Reduced TF Mutated at the Allosteric Cys186-Cys209 Disulfide.

Abstract 1132

Increased TF expression on monocytes contributes to coagulation activation in atherothrombosis, sepsis and cancer. Although it is well established that TF is predominantly non-coagulant and cryptic on unperturbed cells, the molecular mechanisms underlying TF activation remain intensely debated. Antithymocyte globulin (ATG) is a polyclonal rabbit IgG used in transplantation medicine to prevent organ rejection and graft-versus-host disease, but it can cause thrombocytopenia and low-grade DIC. In this study, we investigated the effect of ATG on TF activation. ATG specifically activated TF on cells of the (myelo)monocytic lineage, including THP1, HL60 and U937 cells, as well as isolated peripheral blood monocytes, but not on myeloma, glioma and epithelial cancer cells. Membrane-bound ATG dose-dependently enhanced the TF-specific procoagulant activity (PCA) of THP1 cells 30 ± 15-fold compared to control IgG, as assessed by one-stage clotting, chromogenic Xa generation assay or thrombin generation measured by the Technothrombin™ TGA. ATG rapidly (within 2 min) activated THP1 cells dependent on the Fc portion, but independent of Fcγ receptors, and required heat-labile plasma components for TF induction. ATG did not further enhance TF activity after cell lysis and had no effect on clotting induced by recombinant TF, demonstrating specific effects on TF activation rather than the coagulation reaction itself. ATG-mediated TF activation on THP1 cells did not involve inhibition of surface-floated TFPI and was more potent than, and synergized with, calcium ionophore treatment. Although ATG variably induced exposure of low to moderate levels of phosphatidylserine (PS) on the plasma membrane, as measured by annexin V-FITC binding (38 ± 25 vs. 8 ± 5% positive cells, ATG vs. control IgG, n=22), robust TF activation was observed even in experiments with marginal PS upregulation. Consequently, there was no correlation between TF PCA and PS exposure on ATG-treated THP1 cells (r=-0.03), indicating that the primary effect of ATG was on the activation of TF itself. Previous studies have implicated protein disulfide isomerase (PDI) and the TF extracellular allosteric Cys186-Cys209 disulfide bond as determinants for TF activation. Consistently, PDI was detectable at low levels on the surface of unstimulated THP1 cells by flow cytometry and ATG-induced TF activation was inhibited by 70 ± 15% (n=9) with RL90, an inhibitory monoclonal antibody to PDI previously shown to have antithrombotic effects in vivo. RL90 did not decrease PS externalization on ATG-activated cells and had no effect when added to THP1 cells after TF activation had occurred, demonstrating that the antibody specifically blocked the TF activation process. In addition, blocking free thiols of THP1 cells with N-ethylmaleimide slightly increased TF activity, but completely abolished further TF activation by ATG, demonstrating that thiol exchange reactions were critically involved. In order to directly probe the involvement of the allosteric TF disulfide in ATG-mediated TF activation, we raised monoclonal antibodies to Cys186-Cys209 disulfide bond mutants of human TF. Panreactive and mutant TF-specific antibodies were identified. Monoclonal antibody 3D10 did not bind to oxidized recombinant soluble TF, but selectively recognized the Cys186Ala recombinant mutant that retained a free Cys209 prone to S-nitrosylation/glutathionation. Antibody 3D10 completely inhibited ATG-induced TF activation on THP1 cells, but was without effect when added to THP1 cells after TF activation had occurred. The inhibitory effect of 3D10 on TF activation was dose-dependent (5-50 μg/ml), did not require Fcγ receptor binding and did not involve competition with ATG binding to THP1 cells. Preincubation of THP1 cells with 3D10-derived Fab fragments also inhibited ATG-mediated TF activation by 68 ± 7% (n=8) without affecting PS exposure. Importantly, the combination of RL90 and 3D10 Fab had no additive effects, suggesting that both interfered with the same molecular activation pathway. In summary, these data provide novel evidence that TF cellular activation involves intermediates with an altered conformation of the allosteric Cys186-Cys209 disulfide in the TF extracellular domain and further substantiate the concept that TF decryption involves PDI-dependent pathways that are distinct from the exposure of procoagulant PS on cell surfaces.