Binding of Quinine-Dependent Monoclonal Antibodies to Gpiib/Iiia Appears to Involve Drug-Induced Modulation of Antibody Affinity

A major type of drug-induced immune thrombocytopenia (DITP), characterized by an acute, sometimes life-threatening drop in platelets following drug exposure, is caused by a unique type of antibody that recognizes its target on a platelet membrane glycoprotein, usually αIIb/β3 integrin [glycoprotein (GP) IIb/IIIa], only when the sensitizing drug is present in soluble form. Quinine (Qn) is the prototypic drug that causes this complication, but many other drugs have been implicated. It is widely thought that drug-dependent, platelet-reactive antibodies (DDAbs) characteristic of DITP recognize drug-induced structural modifications of platelet glycoproteins (GP), but this has not been confirmed experimentally. The mechanism responsible for DDAb binding is difficult to study with human DDAbs, which are often poly-specific and in short supply. We used newly-developed, Qn-dependent monoclonal antibodies (IgG₁ mAbs 314.1, 314.3) that recognize the N-terminus of GPIIb and closely mimic the serologic behavior of antibodies from patients with Qn-induced immune thrombocytopenia (Blood 2009; 113;1105-11) as an alternative tool for studying the molecular basis of drug-dependent antibody binding.

Previous studies failed to demonstrate a docking site for Qn in domains of GPIIb/IIIa that are known targets for the "314" mAbs and for human Qn-dependent antibodies. Therefore we examined an alternative possibility - that binding of drug to antibody might be the first step in DDAb binding. For this purpose, Qn was perfused over the "314" mAbs immobilized on Biacore chips and surface plasmon resonance (SPR) signals were recorded. Findings showed that Qn binds specifically to both mAbs with high affinity (Kd about 30 nM) and with 2:1 stoichiometry (Qn to mAb), consistent with recognition of Qn by complementarity determining regions (CDR) of the mAbs. To characterize monovalent binding of mAb to GPIIb/IIIa, purified integrin in 0.1% triton X-100 was perfused over immobilized mAb 314.1 and SPR signals recorded. Weak, but specific binding was observed in the absence of Qn (Kd 11 uM) that was enhanced 5-fold (Kd 2.2 uM) when Qn was present. Kds for Qn-dependent binding of mAb 314.1 (bivalent interaction) and its Fab fragment (monovalent interaction) to GPIIb/IIIa were determined by flow cytometry using labeled antibody and Fab under conditions that did not require washing prior to direct measurement of platelet bound IgG and Fab. Weak Fab binding was observed in the presence of Qn (≈19 uM) but with intact IgG the effective Kd was reduced to 0.15 nM, reflecting a 100,000-fold increase in avidity.

Together with studies that have failed to demonstrate any docking site for Qn on GPIIb/IIIa, the findings support a model in which DDAb-GPIIb/IIIa interaction starts with binding of drug to the antibody CDR, leading to a structural change that markedly increases the avidity of antibody for a weak autoantigen. A requirement for bivalent antibody-target interaction to achieve tight binding could explain why DDAbs almost invariably recognize GPIIb/IIIa or GPIb/IX, the most highly expressed platelet glycoproteins. How this type of DDAb is induced by drug remains uncertain but the findings are consistent with a model in which sensitization starts with drug-induced modification of a B cell receptor that increases its affinity for a weak autoantigen.