Structural and Functional Characterization of a Single Chain β3 Integrin in Solution and on the Cell Surface and the Implications on Integrin Activation and Pathogenesis of Anti-HPA-1a Alloantibodies

The β3 integrins, including αIIbβ3 and αVβ3, are among the most well studied integrins. αIIbβ3 is specific to platelets and essential for hemostasis and thrombosis. αVβ3 is ubiquitously expressed in many cell types, including platelets and endothelial cells, and is important for cell growth and survival. The β3 subunit contains the headpiece formed by PSI, hybrid, βI, and I-EGF1 domains, the leg piece formed by the I-EGF 2-4 and β-tail domains, a transmembrane (TM) domain and a short cytoplasmic tail (CT). Structural studies revealed both αIIbβ3 and αVβ3 complexes as a bent conformation with the headpiece folded onto the leg piece, representing a resting state. Upon activating by ligand binding to the headpiece or talin/kindlin binding to the CT, β3 integrin undergoes a large conformational rearrangement, showing the headpiece extension from the leg piece and the separation of the leg from the α subunit. Only the interface at the headpiece between the β3 βI and the α β-propeller domains is maintained during the bent to extended transition. The interactions at the α/β TM, CT and leg domains are broken in the active conformation. It remains unknown how the extended state of β3 is triggered and maintained, especially whether the α subunit contributes to regulating the β3 conformation.

A Leu₃₃Pro polymorphism at the β₃ PSI domain, forming human platelet antigen-1a/1b (HPA-1a/1b), can induce alloantibodies against incompatible HPA-1a during pregnancy or after blood transfusion, leading to two severe bleeding disorders, fetal/neonatal alloimmune thrombocytopenia (FNAIT) and post-transfusion purpura (PTP). The severe risk of FNAIT is intracranial hemorrhage (ICH), often leading to death or life-long neurological disability. The direct cause of ICH remains unknown, but recent studies suggest a strong connection between the ICH and the presence of a subtype of anti-HPA-1a that is specific to the αVβ3 complex on endothelial cells. The other anti-HPA-1a subtypes are αIIbβ3 specific and complex independent. The αVβ3-specific anti-HPA-1a can interfere with endothelial cell function by inhibiting the αVβ3-mediated cell adhesion and angiogenesis, which may be responsible for the development of ICH. Certain anti-HPA-1a antibodies were also found to affect αIIbβ3-mediated platelet function. It is unknown how the anti-HPA-1a antibodies, which bind to the β3 PSI domain, can distinguish the β3 between αVβ3 and αIIbβ3, and how they block β3 integrin function.

To address the above questions, we engineered a chimeric β3 integrin (cβ3) by replacing the βI domain with the αI domain of αL integrin, enabling to express the full-length β3 as a monomer in the absence of α subunit. With this strategy, the conformation of β3 can be studied without α subunit. On the cell surface, the cβ3 constitutively adopts an extended active conformation, evidenced by the binding of the αL ligand ICAM-1 in a metal ion-dependent manner, and the binding of active conformation-dependent anti-β3 mAbs. Very surprisingly, the cβ3 ectodomain exists in a compact conformation in solution as indicated by size exclusion chromatography and dynamic light scattering. This was confirmed by a high-resolution crystal structure of cβ3, which shows a bent shape that is comparable with the β3 conformation in the bent structures of αVβ3 and αIIbβ3. Our cβ3 structure revealed a rotational movement at the ankle region connecting the I-EGF4 and the β-tail domains. Mutagenesis study demonstrated that the plasticity of the ankle region regulates the activation of β3 integrin. These data demonstrate that the β3 integrin can adopt either a bent or extended conformation even without α subunit. Our data suggest that the TM, CT, cell membrane, and/or other interacting molecules are responsible for the active form of β3 on the cell surface. We next used our monomeric cβ3 to detect the anti-HPA-1a antibodies in FNAIT and PTP, aiming to distinguish the complex-dependent and independent subtypes. Remarkably, we found that anti-αIIbβ3 alloantibodies in FNAIT only recognize the bent conformation of either αIIbβ3 or cβ3 but not the extended form. Structural analysis of the PSI domains of the available β3 crystal structures revealed conformational diversity and flexibility, which may account for the heterogeneity of the anti-HPA-1a antibodies. More importantly, the bent conformation specific anti-HPA-1a antibodies may stabilize the resting integrin and block its cellular function.