The anti-phospholipid syndrome refers to a range of autoimmune conditions which are characterised by venous or arterial thrombosis, recurrent strokes, pulmonary embolism, recurrent pregnancy loss or obstetric complications and the presence of circulating antibodies with specificity to a range of phospholipids; including phosphatidylserine and cardiolipin. The syndrome is the leading cause of vascular thrombosis in children. It sometimes accompanies other autoimmune conditions such as systemic lupus erythematosus (SLE).
Recently it has become apparent that at least some of the auto-antibodies are in fact directed against phospholipid-protein complexes or against the lipid-binding proteins themselves: ^2-glycoprotein I (^2 -GPI) and prothrombin being the major targets. ^2 -GPI is a circulating plasma protein which binds to artherogenic-lipoprotein (Lp(a)) which is structurally homologous to plasminogen. It has been suggested that it could form a ternary complex with Glu-plasminogen and tissue-associated plasminogen activator, facilitating plasmin activation. As described already, endothelial surface-bound annexin 2/S100A10 complex functions as a 'fibrinolytic-receptor' by binding plasminogen and tPA. It has also been shown that annexin 2 forms a complex with ^2-GPI on the surface of endothelial cells (Ma et al., 2000), allowing the latter protein to bind and generate an anti-coagulant environment at this surface.
Antibodies directed against ^2 -GPI further enrich the protein on the cell surface but these promote a p38 Map-kinase signalling cascade which results in increased expression of tissue factor (TF) and reduced expression of thrombomodulin on the surface of cells. TF is a major initiator of coagulation and increased levels of TF expression have been measured on endothelial cells treated with anti-phospholipid antibodies and on monocytes both ex vivo and in vivo (Yasuda et al., 2005; Lopez-lira et al., 2005, Lopez-Pedrera et al., 2006). Thrombomodulin is a potent anti-coagulant protein which limits activation of thrombin, so the net result of circulating anti-phospholipid antibodies is to usurp the anti-coagulative, protective mechanism and initiate a pro-coagulation cascade.
Patients with APL have increased circulating titres of anti-annexin 2 antibodies (Cesarman-Maus et al., 2006). Anti-annexin 2 IgG enhances the expression of tissue factor on endothelial cells more than 6 fold, significantly blocks plasmin generation in a tPA-dependent generation assay (independently of beta-2-glycoprotein-I), and significantly inhibited cell surface plasmin generation on human umbilical vein endothelial cells. The exact interplay between ^2-GPI and annexin 2 requires careful dissection but the importance of annexin 2 in the regulation of the coagulability of the environment is clear.
Despite the strength of the evidence for annexin 2 involvement in the development of cellular pathologies, Rand's initial description of the annexinopathies and APL syndrome in particular was not concerned with annexin 2, but rather with annexin 5. The surface availability of anionic charged phospholipids is critical for the initiation and maintenance of thrombosis. Annexin 5 prevents formation of prothrombinase and tenase complexes (intrinsic pro-coagulation complexes formed of factor Villa and factor IXa) by competing for these sites (Andree et al., 1992). It has also been shown to significantly displace preadsorbed anticardiolipin-ß-GPI complexes from such membranes and to modify platelet aggregation and accretion into larger complexes under the conditions of blood flow (van Heerde et al., 1994). Significantly annexin 5 also attenuates the nucleating effects of core debris which can cause additional clots when they are released from ruptured plaques. A polymorphism in the Kozak sequence of the annexin 5 gene has been discovered that does not change the sequence of the protein but has been shown to correlate with an increase in the level of annexin 5 expression and abundance of the protein in the circulation. It has been suggested that this gives some protection against thromboembolic disease (González-Conejero et al., 2002), though this has not been confirmed by other groups (van Heerde et al, 2003).
Annexin 5 is expressed at high levels on the surface of syncytiotrophoblasts where its anti-coagulant activity is thought to protect the developing foetus. Annexin 5 has been proposed to form a 'shield' over the surface of the placenta, adopting a tessellating array on the exposed phospholipids and preventing the circulating pro-clotting factors from finding purchase. In a mouse model in which labelled annexin 5 was injected into Balb-c mice it accumulated on numerous foci on the apical surface of the trophoblasts. Antibodies with specificity to annexin 5 (which binds phosphatidylserine) are commonly associated with APL and it has been suggested that these antibodies or indeed anti-phospholipid antibodies could interfere with annexin 5 shielding of the trophoblast (Rand et al; 2003; 2004; Rand and Wu 2004; Rand et al., 2005, reviewed in Esposito et al., 2005). When the annexin 5 'shield' was disrupted in pregnant mice by the injection of annexin 5 antibodies, thrombosis and necrosis were seen on the foetal component of the placenta, and various degrees of foetal readsorption were observed (Wang et al., 1999).
The significance of anti-annexin 5 antibodies to the pathogenesis of this disease is not entirely uncontentious. A thrombotic association may only occur when anti-phospholipid antibodies are also present, so they may not be a useful independent marker of risk (Arnold et al., 2001). In this moderately sized study, there was not a single patient with antibodies against both annexin 5 and ß2-GPI. This could suggest that there are two, independent routes to the disease: a ß2-GPI/annexin 2 path and an aPL/annexin 5 path. A clear distinction between anti-ß2-GPI and anti-annexin 5 was found in regard to late recurrent miscarriage (Zammiti et al., 2006) and no correlation was found between the presence of anti-annexin 5 antibodies and thrombosis in APL (de Laat et al., 2006). These authors do however find a positive correlation with the mutation in the Kozak sequence described above.
It is clear that there are still issues that need to be resolved in this field. Methodological differences have been highlighted by some of the authors to explain their differing conclusions, but there may be some specific property of annexins that renders them susceptible to this kind of analytical quandary. In various other autoimmune conditions it has been suggested that autoantibodies develop as a result of antigen creep from the so-called 'natural antibodies' with broad specificity for a range of common pathogens. Anti-annexin antibodies have been discovered in a wide range of autoimmune conditions; perhaps the charged, helix-rich structure of annexins, in complex with charged lipids just happens to represent a surface which can easily mimic antigens derived from pathogens. Perhaps the association of annexins with membrane surfaces, in inflammatory environments and on phago-somal and endosomal membranes (as will be discussed later) means that they feature regularly as the 'usual suspects' in immunological identity parades. That annexin-specific antibodies have the potential to be pathogenic is clear; whether they will be useful markers for autoimmune conditions is possible, and whether or not they are of clinical relevance still remains to be proven beyond reasonable doubt.
Was this article helpful?