DIFFERENTIATING BETWEEN HEPARIN-INDUCED THROMBOCYTOPENIA AND VACCINE-INDUCED THROMBOTIC THROMBOCYTOPENIA ANTIBODY EPITOPES ON PF4

Introduction: Heparin-induced thrombocytopenia (HIT) is a potentially life-threatening adverse event that occurs in approximately 5% of patients following heparin administration to prevent clotting during cardiac as well as orthopedic surgeries, kidney dialyses and blood transfusions. HIT is mediated by antibodies that bind complexes of platelet factor 4 (PF4 or CXCL4) and heparin that cross-link Fc receptors on platelets to cause platelet activation, thrombocytopenia with or without thrombosis, potentially leading to death. Recently, a new anti-PF4 disorder has emerged in response to COVID-19 adenoviral-vector based vaccines against SARS-CoV-2 known as vaccine-induced immune thrombotic thrombocytopenia (VITT). VITT is a rare but serious adverse syndrome reported 5-30 days after administration of a COVID-19 adenoviral vector vaccine (Oxford-AstraZeneca ChAdOx1 nCoV-19 (AZ) and Johnson & Johnson/Janssen (J&J/Janssen)) and is characterized by thrombocytopenia and thrombosis in unusual locations, including cerebral sinus venous thrombosis (CVST). HIT and VITT are characterized by platelet-activating antibodies that bind to PF4 and often lead to thrombocytopenia and/or thrombosis. There are two types of platelet-activating antibodies that lead to HIT, antibodies that activate platelets only in the presence of heparin (heparin-dependent (HD)) and antibodies that can activate platelets in the absence of heparin (heparin-independent (HI)). In VITT, the two types of platelet-activating antibodies are ones that only activate platelets in the presence of PF4 (PF4-dependent) and antibodies that can activate platelets in the absence of PF4 (PF4-independent). We hypothesize that the binding sites of HIT and VITT antibodies on PF4 differ and that identifying these sites could be relevant to differentiate between both anti-PF4 disorders. Methods: All 70 PF4 mutants produced by alanine-scanning mutagenesis (ASM) were expressed and purified from E. coli BL21 (DE3) cells. ASM was performed using HIT (n=10) and VITT (n=14) patient samples to determine the binding sites of HIT and VITT antibodies on PF4. Clinical outcomes of CVST, DVT, PE incidences in VITT patients with PF4-dependent and PF4-independent antibodies were also analyzed. In addition, ASM was performed on the HIT-mimicking monoclonal antibodies (1E12, 1C12, 2E1, P4B1, P3D4, P3E10) to confirm the pathogenic antibody binding sites of HIT antibodies on PF4. Finally, all 70 PF4 mutants were screened against HIT (n=22), VITT (n=9) and non-HIT (n=12) patient samples and we identified 10 PF4 mutants that differentiate between HIT and VITT using an EIA. Results: We confirmed all 70 PF4 mutants were free of contaminants compared to bacterial lysate by visualizing the purity using gel electrophoresis. Using a cut-off of 50% or greater reduction in antibody binding to PF4 mutants compared to wild-type PF4, we determined that most monoclonal antibodies that mimic platelet-activating HIT antibodies, as well as antibodies purified from HIT patients bind to a site on PF4 like the KKO site, a known monoclonal antibody against PF4/heparin complexes that mimics platelet-activating HIT antibodies. While most VITT antibodies recognize a site on PF4 overlapping the heparin-binding site, we found a subset of VITT patients have antibodies that recognize an additional site on PF4 that is like the KKO site. By observing the incidences of clinical thrombotic characteristics in VITT patients, it was determined that CVST (6/14, 42.9%), DVT (4/14, 28.6%), and PE (4/14, 28.6%) were the most common types of thrombosis. This unusually high incidence of VITT is not typically seen in HIT. Information on the binding characteristics of HIT and VITT anti-PF4 antibodies have sparked interest on whether there are characteristics that can help differentiate between HIT and VITT. After screening the 70 PF4 mutants against HIT and VITT patient samples, we then identified 10 PF4 mutants (H23A, R20A, T25A, E28A, K46A, R49A, K50A, K62A, K65A, K66A) that help differentiate between HIT and VITT in an EIA. We also screened the 70 PF4 mutants against non-HIT patients, identified as individuals that have non-platelet activating anti-PF4/heparin antibodies. We found that 5 PF4 mutants (K14A, H23A, V29A, T38A, K50A, K66A) help differentiate HIT from non-HIT in an EIA. Conclusion: Purifying PF4 mutants reduced background reactivity likely caused by contaminants in assays using bacterial lysate. ASM using PF4 mutants revealed that there are 2 distinct binding sites on PF4 that are recognized by either HI and HD HIT or anti-PF4 VITT antibodies. We confirmed that the molecular characteristics of antibodies from VITT and HIT patients dictate the platelet activation profiles in these disorders and can be used to separate VITT from HIT antibodies in the EIA. Thesis Master of Science (MSc)