Characterization of 5B9, a Chimeric Monoclonal Anti-PF4/H Antibody with a Human Fc Fragment and Which Mimics the Effects of HIT Human Antibodies

Heparin-induced thrombocytopenia (HIT) is a frequent drug-adverse event caused in the majority of patients by platelet-activating antibodies (Abs) directed against complexes of heparin (H) bound to platelet factor 4 (PF4). In most cases, HIT Abs are IgG which are potentially pathogenic as they are able to activate platelets directly in the presence of heparin via FcgRIIA receptors. The diagnosis of HIT is based on both clinical and biological criteria. However, despite recent improvements in HIT laboratory assays, a standard is always lacking for both immunological and functional assays. First, platelets from healthy donors exhibit wide variability in their response to HIT antibodies. Second, many patients who synthesize significant levels of antibodies to PF4/H while being treated with heparin do not develop HIT and the mechanisms that regulate the pathogenicity of HIT antibodies have not been fully defined. However, several studies suggested that epitope specificity of IgG antibodies is critical for the pathophysiology of HIT, especially by influencing the stability of PF4 tetramers.

We therefore aimed to develop a monoclonal anti-PF4/H antibody with a human Fc fragment using transgenic mice homozygous for Cg gene of a G-class human immunoglobulin and that directly produce chimeric IgG antibodies.

After immunization, several clones were found to synthesize anti-PF4 IgG antibodies but only one (5B9) produced an IgG1 that specifically bind PF4/H complexes without reactivity against PF4 alone. 5B9 was able to induce dose-dependent platelet activation and aggregation in the presence of low concentrations of UFH (0.1 and 0.5 IU/mL), and this effect was not observed without UFH or with high concentration of UFH (10 IU/ml), and was fully inhibited by IV.3 antibody. In addition, 5B9 with UFH (0.5 IU/mL) induced strong TF mRNA synthesis in isolated monocytes. Injection of 5B9 with UFH to transgenic mice expressing human PF4 and FcgRIIA receptors was followed by significant thrombocytopenia, similar to that observed with KKO, a murine IgG2b anti-PF4/H antibody. Plasma levels of thrombin/anti-thrombin (TAT) also significantly increased after injection of heparin in all mice having received 5B9 or KKO, but this effect was more pronounced on day 1 in 5B9-treated mice (mean value= 47 vs. 5 ng/mL at day 0) than in mice injected with KKO (33 vs. 11 ng/mL, respectively).

Competitive immunoassays were also developed and 15 of 25 plasma samples (60%) collected in HIT patients were shown to inhibit (by at least 20%) the binding of 5B9 to PF4/H complexes compared to 3/25 (12%) samples containing non pathogenic anti-PF4/H antibodies (obtained in patients without HIT). Similar experiments were performed with KKO and its binding to PF4/H was also inhibited by 32% of HIT plasma samples (8/25), and none of the non-HIT plasmas. However, the levels of 5B9 and KKO binding inhibition to PF4/H were highly correlated (R²=0.85), thereby suggesting that the epitopes recognized by the antibodies are similar. Noticeably, KKO was also shown to inhibit in a concentration-dependent manner the binding of 5B9 to PF4/H complexes, further supporting this hypothesis.

5B9 was then sequenced and a docking model was elaborated based on V_H and V_L sequences of 5B9 obtained and a recently described crystal structure of KKO/PF4 tetramer complex. According to this model, 5B9 Fab likely interacts with the B and D monomers of PF4, thus possibly contributing to the stability of tetramers. In addition, the binding of 5B9 to PF4 involves 28 aa, 16 in the B monomer including Asp-7, Gln-9, Pro-34 and Pro-37 and 12 in the D monomer including Gln-9. Importantly, 13 of these 28 aa have also been identified as critical in the formation of PF4/KKO complex (Cai et al, Nat Commun. 2015;6:8277). Three regions (Asp-7 to Thr-15 and Ala-32 to Lys-41 in the B monomer and Gln-9 to Asp-18 in the D monomer) therefore appear particularly important in the binding of both 5B9 and KKO on PF4 tetramers.

In conclusion, 5B9 is the first anti-PF4/H monoclonal antibody that has a human Fc fragment, and fully mimics the effect of HIT human antibodies. 5B9 could therefore be used as a standard in HIT laboratory assays. Moreover, our results support that three regions both recognized by 5B9 and KKO within PF4 tetramers are critical for the binding and pathogenicity of HIT antibodies.