Atomic Level Description of the Immune Complex That Causes Heparin-Induced Thrombocytopenia (HIT)

Heparin-induced thrombocytopenia (HIT) is thrombotic disorder caused by immune complexes containing antibodies to an antigen composed of platelet factor 4 (PF4) and heparin or cellular glycosaminoglycans (GAGs). The structure of these immune complexes and how their composition might contribute to the difference between pathogenic and non-pathogenic anti-PF4 antibodies are unknown. To address these questions, we solved the crystal structures of human recombinant PF4 in complex with Fabs derived from KKO (a murine monoclonal HIT-like antibody that competes with pathogenic human HIT antibodies) and RTO (an isotype-matched non-HIT anti-PF4 antibody) combined with the crystal structure of PF4 complexed with the heparin-mimic pentasaccharide fondaparinux as a model sugar. The PF4 tetramer is asymmetric and is capable of accommodating only two fondaparinux molecules. Fondaparinux binds between monomers A, B and C or between monomers A, C, and D, which stabilizes the AB/CD and AC/BD associations and the resultant tetramer. KKO-Fab binds to the PF4 tetramer by making contacts with now identified residues within each of three PF4 monomers, indicating that tetramerization of PF4 is a critical initiating step in antigen formation. Mutations in the putative KKO epitopes in PF4 abolished antibody binding.Unexpectedly, RTO-Fab binds to the PF4 monomer between the AB dimer interface. Importantly, the amino acid sequence recognized by RTO and KKO show considerable overlap. However, the epitope for RTO is obscured upon tetramer formation, in direct contrast to binding of KKO, which requires tetramer formation to bind. Binding of RTO to the PF4 monomer prevents formation of AB dimers and subsequent tetramerization. In support of these findings, preincubation of PF4 with RTO inhibits KKO induced platelet activation and platelet aggregation in vitro. Based on the analyses of crystal lattices, we propose a new model of the heparin/PF4 complex, in which PF4 tetramers cluster around a semi-rigid linear heparin subunit. Clustering of PF4 on heparin might be required for apposition of sufficient HIT antibodies to induce persistent activation of cellular FcgIIA receptors. Heparin and pathogenic HIT antibodies collaborate to stabilize the ternary immune complex, which leads to the disappearance of binding sites for at least some non-pathogenic HIT antibodies. The balance between anti-monomer and anti-tetramer PF4 antibodies may help determine the probability of clinical disease. This model also helps to explain why RTO-like anti-PF4 antibodies are found so commonly in asymptomatic patients exposed to heparin and why fondaparinux may be antigenic but rarely causes HIT, whereas longer heparin fragments and GAGs extend and render the holo-complex more stable and thereby foster the formation of pathogenic immune complexes. In summary, these crystallographic studies lead to a new model to explain the formation of pathogenic immune complexes that lead to HIT. The inhibitory effect of the anti-PF4 antibody RTO provides a structural basis for the development of new diagnostics and non-anticoagulant therapeutics.