Are posttranslational modifications of β2-glycoprotein I markers for thrombotic risk? Are they triggers of autoimmunity?


For over 30 years, we have known that thrombosis and fetal loss are the signature features of the antiphospholipid syndrome (APS). We understand the major functional components of this syndrome: antibodies, primarily of IgG class, bind to an autoantigen, in many cases β2-glycoprotein I (β2GPI), that itself binds either to negatively charged phospholipids or, more likely, to cell surface receptors such as apolipoprotein E receptor 2 (Apo ER2). These events initiate a chain of intracellular or extracellular signals that lead to thrombosis (1–5). Although the chemistry and cell biology are familiar, it has not been possible to predict in whom and at what point in time patients with antiphospholipid antibodies will have clinical events.

Multiple studies in animals and humans have elucidated elements of the pathophysiology that leads to a diagnosis of clinical APS. The hypotheses tested in the models focus on specific events from the first appearance of antibody to thrombosis, including phenotypic changes in vulnerable tissue, antibody-triggered pathways of inflammation and injury, alterations in coagulation proteins and platelets, cellular activation by stimulation of surface receptors, environmental insults, and coexistent genetic thrombotic states.

The most robust current theories of pathogenesis (Table 1) incorporate fundamental observations: some form of cell activation occurs, exposing negatively charged phospholipids on the cell surface or initiating cell adhesion or thrombosis; β2GPI is involved in the thrombotic process, even though antibody to β2GPI is not a strong marker of risk; and triggering factors provoke clinical events (6, 7). Early involvement of inflammation, particularly complement activation, has been established in animal models but is less clear in humans.

Table 1. Theories of how antiphospholipid antibodies cause the antiphospholipid syndrome
Aspects of antiphospholipid antibodies that may contribute to pathogenicity
 Features of the autoantibody: class, subclass, or function that predispose to thrombosis
 Features of the antigen recognized: type of phospholipid or coagulation factor
 Exposure of neoepitope in β2-glycoprotein I associated with conformational changes
 Signaling via endothelial surface receptors, including apolipoprotein E receptor 2, annexin 2, Toll-like receptors
 Activation of endothelial cells or platelets, including expression of tissue factor, or adhesion molecules, and inhibition of endothelial cell nitric oxide synthase
 Initiation of inflammation by targeted antibody, including complement activation and release of proinflammatory cytokines
 Increased vulnerability of endothelium triggered by oxidative stress, oral contraceptives, surgery, infections, smoking, etc.
 Other prothrombotic abnormalities, including methylenetetrahydrofolate reductase, factor V Leiden, etc.

In this issue of Arthritis & Rheumatism, Ioannou and colleagues add to a body of work that secures for β2GPI a central role in the pathophysiology of APS and provides evidence that alterations in β2GPI are associated with increased risk of thrombosis (8). Beta-2-glycoprotein I (also known as apolipoprotein H), an incompletely understood protein with incompletely defined physiologic function, circulates in large quantity in plasma. Subjects deficient in β2GPI appear to be healthy, compounding the mystery of its role.

A part of the complement control protein family, β2GPI consists of 5 “sushi domains” (wrapped like pieces of sushi). The molecule is J-shaped. Its hook end, the fifth domain, is the site of its attachment to the cell surface. The highly positively charged fifth domain binds to negatively charged surfaces, such as the phosphatidylserine that translocates from the inner cell membrane to the surface in activated or apoptotic cells. Recent studies have indicated that the fifth domain of β2GPI also binds to Apo ER2 receptors and thereby activates platelets and endothelial cells (1, 2, 7). Upon binding, β2GPI undergoes a conformational change that likely explains its transformation from a benign circulating protein to an active participant in thrombosis (4). While the fifth domain targets β2GPI to cell surfaces, the first domain contains a neoepitope, exposed after the molecule binds to a surface, that is recognized by antiphospholipid antibodies. Binding and crosslinking of the first domain by antibodies trigger intracellular and extracellular pathways that result in thrombosis and may cause pregnancy complications.

In the current study (8), Ioannou and colleagues investigated whether posttranslational modification of β2GPI, independent of binding-induced conformational change, correlates with thrombosis. They posed the question through a clinical study, examining total β2GPI and the proportion of oxidized β2GPI (with oxidation defined as lacking free thiols), in a large multicenter study. Patients with nonthrombotic autoimmune disease, patients with thrombosis without autoimmune disease, and healthy individuals served as controls. Pregnancy loss as a manifestation of APS was not considered, although its mechanism may (or may not) be different. Indeed, most women with APS-related fetal loss do not experience thromboses.

The authors found that APS patients had higher levels of total β2GPI and proportionally more oxidized β2GPI than any of the control groups. In addition, oxidized β2GPI was associated with lupus anticoagulant, an important observation because lupus anticoagulant is the dominant predictor of thrombotic clinical events in APS. They propose that relative levels of oxidized β2GPI can be used to stratify patients by thrombosis risk. However, they were not able to develop a more complete risk profile because the study, which included 139 APS patients, did not address the effects of systemic lupus erythematosus, anticoagulation treatment, clinical manifestations of APS, or risk associated with a decrease in reduced β2GPI. The authors speculate that clinical states characterized by increased oxidative stress, such as pregnancy and infection, lead to elevated levels of oxidized β2GPI, which is more immunogenic and likely to break tolerance, and that the relative decrease in reduced β2GPI leads to increased thrombosis risk. Importantly, in vitro studies from this group provide evidence that reduced β2GPI protects endothelial cells against oxidative injury and maintains their viability (9). These studies identify a new and pivotal role for β2GPI in the induction of thrombosis in APS.

The work of Ioannou et al demonstrates clear biochemical differences between normal and pathogenic β2GPI, links these differences to clinical phenotype, and suggests posttranslational mechanisms by which β2GPI can be modified to become pathogenic. Elucidating the structural biology of β2GPI is one piece of the puzzle that we now know has broad implications. To understand pathophysiology and use relative levels of the posttranslationally modified forms of β2GPI for risk stratification, one also must consider antiphospholipid antibody class, antigen localization, activation of effector cells, initiation of pathways of inflammation (complement and cytokines) and coagulation, and environmental triggers.

A number of questions about antiphospholipid syndrome cannot be answered by the observations of Ioannou et al. Why do some patients with high-titer antiphospholipid antibodies have no illness? How does oxidation of β2GPI relate to familial and ethnic skewing in APS? Is β2GPI more antigenic in its oxidized state? Does oxidized β2GPI explain the basis for association of antiphospholipid antibodies with systemic lupus erythematosus? Is there a more effective treatment that blocks upstream events that will supplant anticoagulation?

Larger studies with longitudinal followup must confirm and extend the exciting new findings reported by Ioannou and colleagues. The onus is on the international community of physicians and scientists who care for and study APS patients to answer these questions.


Drs. Lockshin and Salmon drafted the article, revised it critically for important intellectual content, and approved the final version to be published.