Determinants of PF4/Heparin Immunogenicity in a Murine Model of HIT.

Heparin-Induced Thrombocytopenia (HIT) is a life-threatening immune-mediated reaction caused by platelet factor 4 (PF4)/heparin complexes. In recent in vitro studies, we have shown that human and monoclonal HIT antibodies recognize and optimally bind to ultra-large complexes (ULC; MW>670kDa) formed at equimolar ratios of PF4 and heparin. We undertook the following studies to determine if the in vivo immunogenicity of PF4/heparin complexes showed similar requirements for stoichiometric molar ratios of PF4:heparin (PHR). Using a previously described murine immunization model, mice were immunized with antigen by retro-orbital injection daily for 5 days without adjuvant. Blood was collected at baseline and at weekly intervals for 4–6 weeks. Plasma was tested for antibody development by a murine PF4/heparin (mP+H) ELISA. To determine the effects of PHR on mP+H antibody production in vivo, PF4 was mixed with unfractionated heparin (UFH) at molar ratios of 2:1, 1:1 or 1:2 and injected into mice (n=5/cohort). Antibody levels (mean A_450nm ± SD) of mice injected with PHR 2:1 (0.79 ± 0.77) were higher than mice injected with PHR 1:1 (0.23± 0. 2), or PHR 1:2 (0.09 ± 0.02, p<0.05 for PHR 2:1 v. PHR 1:2). Because in vitro studies showed that UFH was far more capable of generating ULCs when mixed with PF4 than low-molecular weight heparin (LMWH) or fondaparinux, mice (n=5/cohort) were injected with buffer, or mPF4 mixed with UFH, or enoxaparin (a LMWH) or the pentasaccharide, fondaparinux at doses that are therapeutically equivalent to doses given to humans. Antibody levels (mean A_450nm ± SD) were significantly higher in mice injected with UFH (3.340 + 0.15) than in mice injected with buffer (0.0451 ± 0.013, p<0.001 for buffer v UFH), LMWH (0.8526 ± 0.44, p<0.001 for LMWH v UFH) or the pentasaccharide (0.80 ± 0.0.7, p<0.001 for pentasaccharide v UFH). Because PF4/heparin ULCs are composed of repeating structural antigenic units, we next investigated the role of toll-like receptors (TLRs), a family of pattern recognition receptors (PRRs) that recognize repeating structural determinants on microbes and foreign antigens. To determine if mP+H complexes engage TLRs, we injected mP+H into wild-type (WT, n=5) or mice deficient in MyD88 (n=7), an adaptor molecule critical for intracellular signaling by most TLRs. Antibody levels (mean A_450nm ± SD) did not significantly differ between WT (1.5 ± 0.6) and MyD88 null mice (0.9 ± 0.5). In summary, these studies indicate that in vivo antibody responses to mP+H complexes are critically dependent on stoichiometric ratios favoring ULC formation and that immune activation by PF4/heparin is independent of MyD88 signaling pathways.